A method and device for analyzing the results of a metrological verification device

Abstract

The present invention discloses a method and device for analyzing the results of a metrological verification device. The verification device is verified in real time by using an error distribution function model of an electric energy meter to obtain multiple error measurement values. A time series is constructed based on each error measurement value and the verification date corresponding to each error measurement value. Then, a time series is constructed according to the verification results, and the least squares method is used to fit the time series. In particular, the possibility of insignificant characteristic estimation coefficients is used to determine the real-time verification results, so as to determine whether there is a linear change relationship between the verification results and the dates. The time series trend warning formula is used to alarm the trends of the error, stability, and repeatability of the verification device respectively. By using the error distribution function model of the electric energy meter to verify the verification device in real time, and after obtaining the verification results, the verification results are analyzed and judged to obtain warning information, the verification efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent electricity meters, and in particular to a method and device for analyzing the results of a metrological verification device. Background Art

[0002] With the wide application of intelligent electricity meters and the rapid growth of the business volume of intelligent meter verification, using an automated verification pipeline to verify intelligent electricity meters has become the main verification method for intelligent electricity meters. The automated verification pipeline features a highly concentrated distribution of verification devices, a small number of verification personnel, and a greatly improved verification efficiency. In the current automated verification mode, the number of verification devices is huge, the operation efficiency of the verification system is high, and the daily verification quantity is in the tens of thousands. Therefore, ensuring that the standard performance of the intelligent electricity meter verification device continuously meets the verification requirements is the focus of current metrological standard management. Implementing intermediate checks can promptly detect the inaccuracy of measured values or take remedial measures in a timely manner after inaccuracy, prevent the use of detection equipment that does not meet the requirements of technical specifications, and reduce or lower the costs and risks generated therefrom. Currently, the intermediate checks of verification devices are mainly carried out manually, including a series of cumbersome and time-consuming tasks such as completing the wiring of the verification standard and the verification device, setting and operating related equipment, exporting and processing verification data, and sorting and archiving verification reports.

[0003] The intermediate check frequency of the electricity meter verification device is generally more than 3 times a year. The existing intermediate check methods still rely on manual operation, with low operation efficiency and a great impact on daily verification production. It is difficult to ensure the frequency and timeliness of the check. At the same time, the test process and data processing are greatly affected by human factors.

[0004] In addition, according to the current check frequency, the data in the intermediate check report is seriously insufficient, and the check results are mostly manually recorded. Only a simple conclusion can be given on whether the current check result meets the requirements, and no more valuable analysis can be further made on the check results of the verification device. At the same time, it is difficult to promptly discover the abnormal trend of the verification device from the manual check results, and trend prediction cannot be carried out to reduce or lower the costs and risks caused by the abnormality of the verification device. Summary of the Invention

[0005] To solve the above technical problems, embodiments of the present invention provide a method and device for analyzing the results of a metrological verification device. Through the electricity meter error distribution function model based on the historical verification data of intelligent electricity meters in the same batch, online real-time verification of multiple verification devices is realized, and then a linear regression method is used to analyze the verification results. When it is found that the trend of the verification results is abnormal, a timely reminder is given. Through early determination, the production costs and risks caused by the abnormality of the verification device are reduced, and the verification efficiency and accuracy are improved.

[0006] The first aspect of the embodiment of the present invention provides a method for analyzing the results of a metrological verification device, and the method includes:

[0007] Using the electric energy meter error distribution function model to conduct real-time verification on the verification device to obtain a verification result, wherein the electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model, and the verification result includes multiple error measurement values;

[0008] Sorting according to each error measurement value and the verification date corresponding to each error measurement value to obtain a time series;

[0009] Performing least squares fitting calculation on the time series to obtain a first result, determining whether the first result meets a preset condition, if not, sending a trend warning message for the verification device, after performing sliding window truncation calculation on the time series, obtaining a second result and a third result, determining whether the second result and the third result meet the preset condition, if not, sending a repeatability warning message for the verification device, after performing stability calculation on the time series, obtaining a fourth result and a fifth result, determining whether the fourth result and the fifth result meet the preset condition, if not, sending a stability warning message for the verification device.

[0010] Implementing this embodiment, using the electric energy meter error distribution function model to conduct real-time verification on the verification device to obtain a verification result, wherein the electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model, the verification result includes multiple error measurement values, sorting according to each error measurement value and the verification date corresponding to each error measurement value to obtain a time series, performing least squares fitting calculation on the time series to obtain a first result, determining whether the first result meets a preset condition, if not, sending a trend warning message for the verification device, after performing sliding window truncation calculation on the time series, obtaining a second result and a third result, determining whether the second result and the third result meet the preset condition, if not, sending a repeatability warning message for the verification device, after performing stability calculation on the time series, obtaining a fourth result and a fifth result, determining whether the fourth result and the fifth result meet the preset condition, if not, sending a stability warning message for the verification device. This method improves the verification efficiency and accuracy by using the electric energy meter error distribution function model to conduct real-time verification on the verification device and analyzing and judging the verification result to obtain a warning message.

[0011] In a possible implementation manner of the first aspect, performing least squares fitting calculation on the time series to obtain a first result, determining whether the first result meets a preset condition, if not, sending a trend warning message for the verification device, specifically:

[0012] Perform least squares fitting on the time series to obtain the fitting parameters of slope, intercept, and the probability of insignificance of the estimated coefficient of the verification date;

[0013] When judging whether the probability of insignificance of the fitting parameter slope and the estimated coefficient of the verification date satisfies a preset formula, if it does not satisfy, a verification device trend warning message is issued. The preset formula is:

[0014] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0015] where pvalue represents the probability of insignificance of the estimated coefficient of the verification date, and p accept represents the preset threshold of the probability of insignificance of the estimated coefficient of the verification date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0016] Implementing this embodiment, performing least squares fitting on the time series to obtain the fitting parameters of slope, intercept, and the probability of insignificance of the estimated coefficient of the verification date, and then judging whether the fitting parameter slope and the probability of insignificance of the estimated coefficient of the verification date satisfy the preset conditions. This method uses the least squares fitting method for the time series of the verification result, can dynamically calculate the error change trend, and when the error change trend over time changes significantly, it can timely remind the staff to intervene to prevent the generation of additional production costs and risks.

[0017] In a possible implementation manner of the first aspect, after intercepting the time series with a sliding window, a second result is obtained, and it is judged whether the second result satisfies the preset conditions. If it does not satisfy, a verification device repeatability warning message is issued. Specifically:

[0018] Use the sliding window to intercept the time series to obtain multiple sliding window sequences;

[0019] Calculate the mean value and standard deviation of the error measurement values within each sliding window to obtain the average value and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds the preset value, a verification device repeatability warning message is issued.

[0020] In the implementation of this embodiment, a sliding window is used to intercept the time series to obtain multiple sliding window sequences, and the error measurement values in each sliding window are calculated by mean and standard deviation to obtain the average value and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds the preset value, a repeatability warning information of the calibration device is issued. This method uses the time sliding window method to calculate the mean difference and standard deviation of the measurement results, and then calculates the repeatability of the calibration device. The repeatability of the calibration device can also be assessed to improve the accuracy of the repeatability verification of the calibration device.

[0021] In a possible implementation of the first aspect, after performing a sliding window capture on the time series, a third result is obtained, and it is determined whether the third result meets a preset condition. If not, a repeatability warning message of the calibration device is issued, specifically:

[0022] Perform mean and standard deviation calculations on the error measurement values within each sliding window to obtain the mean and standard deviation of each sliding window;

[0023] After constructing the standard deviation time series using the standard deviation of each sliding window, the standard deviation time series is fitted by the least squares method to obtain the slope and intercept of the fitting parameters of the standard deviation and the probability that the estimated coefficient of the calibration date is not significant. If the slope of the fitting parameters and the probability that the estimated coefficient of the calibration date is not significant do not meet the preset formula, a warning message about the trend of the repeatability measurement result of the calibration device is issued. The preset formula is:

[0024] tr=(p value <p accept )*((b>0)+1)*(abs(b)>b accept )

[0025] Among them, pvalue represents the possibility that the estimated coefficient of the test date is not significant, p accept Indicates the preset threshold value of the possibility that the estimated coefficient of the test date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the slope of the preset fitting parameter.

[0026] In one possible implementation of the first aspect, after performing stability calculation on the time series, a fourth result and a fifth result are obtained, and it is determined whether the fourth result and the fifth result meet a preset condition. If not, a warning message about a trend of repeatability measurement results of the calibration device is issued, specifically:

[0027] Perform mean and standard deviation calculations on the error measurement values within each sliding window to obtain the mean and standard deviation of each sliding window;

[0028] Perform first-order difference calculation using the average value of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than the preset first-order difference preset value, an early warning message for the stability of the verification device is issued;

[0029] Construct a difference time series using the first-order difference results of each sliding window, and perform least squares fitting estimation on the difference time series to obtain the fitting parameter slope, intercept of the first-order difference, and the probability of insignificance of the estimation coefficient of the verification date. If the fitting parameter slope and the probability of insignificance of the estimation coefficient of the verification date do not satisfy the preset formula, an early warning message for the trend of the stability measurement result of the verification device is issued, where the preset formula is:

[0030] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0031] where pvalue represents the probability of insignificance of the estimation coefficient of the verification date, and p accept represents the preset threshold of the probability of insignificance of the estimation coefficient of the verification date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0032] The second aspect of the embodiments of the present invention provides a device for analyzing the results of a metrological verification device, and the device includes:

[0033] A verification module, configured to perform real-time verification on the verification device using an electric energy meter error distribution function model to obtain a verification result, where the electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model, and the verification result includes multiple error measurement values;

[0034] A time series construction module, configured to organize according to each error measurement value and the verification date corresponding to each error measurement value to obtain a time series;

[0035] A judgment module, configured to perform least squares fitting calculation on the time series to obtain a first result, judge whether the first result satisfies a preset condition, if not, issue an early warning message for the trend of the verification device, perform sliding window truncation calculation on the time series to obtain a second result and a third result, judge whether the second result and the third result satisfy the preset condition, if not, issue an early warning message for the repeatability of the verification device, perform stability calculation on the time series to obtain a fourth result and a fifth result, and judge whether the fourth result and the fifth result satisfy the preset condition, if not, issue an early warning message for the stability of the verification device.

[0036] In a possible implementation manner of the second aspect, the judgment module includes a calculation unit and a judgment unit,

[0037] Among them, the calculation unit is used to perform least squares fitting on the time series to obtain the fitting parameters of slope, intercept, and the non-significance probability of the estimated coefficient of the verification date;

[0038] The judgment unit is used to judge whether the non-significance probability of the fitting parameter slope and the estimated coefficient of the verification date satisfies a preset formula. If not, it issues a trend warning message for the verification device. Among them, the preset formula is:

[0039] tr = (p value < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0040] Among them, pvalue represents the non-significance probability of the estimated coefficient of the verification date, p accept represents the preset threshold of the non-significance probability of the estimated coefficient of the verification date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0041] In a possible implementation manner of the second aspect, the judgment module further includes an interception unit and a standard deviation calculation unit.

[0042] Among them, the interception unit is used to intercept the time series by using a sliding window to obtain multiple sliding window sequences;

[0043] The standard deviation calculation unit is used to calculate the mean value and standard deviation of the error measurement values within each sliding window to obtain the average value and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds the preset value, a repeatability warning message for the verification device is issued.

[0044] In a possible implementation manner of the second aspect, after intercepting the time series by using a sliding window, a third result is obtained. It is judged whether the third result meets the preset conditions. If not, a repeatability warning message for the verification device is issued. Specifically:

[0045] Calculate the mean value and standard deviation of the error measurement values within each sliding window to obtain the average value and standard deviation of each sliding window;

[0046] After constructing a standard deviation time series by using the standard deviations of each sliding window, perform least squares fitting calculation on the standard deviation time series to obtain the fitting parameters of slope, intercept, and the non-significance probability of the estimated coefficient of the verification date of the standard deviation. If the fitting parameter slope and the non-significance probability of the estimated coefficient of the verification date do not satisfy the preset formula, a trend warning message for the repeatability measurement result of the verification device is issued. Among them, the preset formula is:

[0047] tr = (pvalue < p accept)*((b>0)+1)*(abs(b)>b accept )

[0048] Among them, pvalue represents the possibility that the estimated coefficient of the test date is not significant, p accept Indicates the preset threshold value of the possibility that the estimated coefficient of the test date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the slope of the preset fitting parameter.

[0049] In one possible implementation of the second aspect, after performing stability calculation on the time series, a fourth result and a fifth result are obtained, and it is determined whether the fourth result and the fifth result meet a preset condition. If not, a warning message about a trend of repeatability measurement results of the calibration device is issued, specifically:

[0050] Perform mean and standard deviation calculations on the error measurement values within each sliding window to obtain the mean and standard deviation of each sliding window;

[0051] Performing first-order difference calculation using the average value of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than a preset first-order difference preset value, a stability warning message of the verification device is issued;

[0052] The differential time series is constructed using the first-order difference results of each sliding window, and the differential time series is fitted and estimated using the least squares method to obtain the slope and intercept of the fitting parameters of the first-order difference and the probability that the estimated coefficient of the calibration date is not significant. If the slope of the fitting parameters and the probability that the estimated coefficient of the calibration date is not significant do not meet the preset formula, a trend warning information of the stability measurement result of the calibration device is issued, where the preset formula is:

[0053] tr=(pvalue<p accept )*((b>0)+1)*(abs(b)>b accept )

[0054] Among them, pvalue represents the possibility that the estimated coefficient of the test date is not significant, p accept Indicates the preset threshold value of the possibility that the estimated coefficient of the test date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the slope of the preset fitting parameter.

[0055] The present invention uses an electric energy meter error distribution function model to perform real-time verification on a calibration device to obtain a verification result, wherein the electric energy meter error distribution function model is constructed by the central limit theorem and the Bayesian hierarchical model, and the verification result includes multiple error measurement values, which are sorted according to each error measurement value and the calibration date corresponding to each error measurement value to obtain a time series, and the time series is fitted by the least squares method to obtain a first result, and it is judged whether the first result meets a preset condition. If not, a calibration device trend warning message is issued, and a sliding window interception calculation is performed on the time series to obtain a second result and a third result, and it is judged whether the second result and the third result meet the preset condition, and if not, a calibration device repeatability warning message is issued, and a stability calculation is performed on the time series to obtain a fourth result and a fifth result, and it is judged whether the fourth result and the fifth result meet the preset condition, and if not, a calibration device stability warning message is issued. The method uses the electric energy meter error distribution function model to perform real-time verification on the calibration device, and after obtaining the verification result, the verification result is analyzed and judged to obtain a warning message, thereby improving verification efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 : A schematic flow chart of an embodiment of a method for analyzing measurement and verification device results provided by the present invention;

[0057] Figure 2 : A schematic diagram of the device structure of another embodiment of the measurement verification device result analysis method provided by the present invention. DETAILED DESCRIPTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0059] Embodiment 1

[0060] Please refer to Figure 1 , which is a flow chart of an embodiment of a method for analyzing measurement verification device results provided by an embodiment of the present invention, including steps S11 to S13, each of which is specifically as follows:

[0061] S11. Use the electric energy meter error distribution function model to perform real-time verification on the calibration device to obtain verification results, wherein the electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model, and the verification results include multiple error measurement values.

[0062] In this embodiment, by using the historical verification data of smart energy meters in the same batch, based on the central limit theorem and the Bayesian hierarchical model, an error distribution function model of the energy meter higher than the accuracy level of the verification device is constructed, and the verification device is verified in real time to obtain the verification result, and the verification result is recorded and saved.

[0063] It should be noted that the verification result is the error measurement value of each verification device.

[0064] As an example of this embodiment, the error measurement values of each verification device calculated by the error distribution function model of the energy meter within 60 days can be selected.

[0065] S12. Organize according to each error measurement value and the corresponding verification date to obtain a time series.

[0066] In this embodiment, according to the verification result, a time series EquipError is formed according to the date and the error measurement value of each verification device. ts .

[0067] S13. Perform least squares fitting calculation on the time series to obtain a first result, and judge whether the first result meets the preset conditions. If not, send a trend warning message for the verification device. After performing sliding window truncation calculation on the time series, obtain a second result and a third result, and judge whether the second result and the third result meet the preset conditions. If not, send a repeatability warning message for the verification device. After performing stability calculation on the time series, obtain a fourth result and a fifth result, and judge whether the fourth result and the fifth result meet the preset conditions. If not, send a stability warning message for the verification device.

[0068] In a preferred embodiment, perform least squares fitting calculation on the time series to obtain a first result, and judge whether the first result meets the preset conditions. If not, send a trend warning message for the verification device. Specifically:

[0069] Perform least squares fitting on the time series to obtain the fitting parameter slope, intercept, and the probability of insignificance of the estimated coefficient of the verification date;

[0070] When judging whether the fitting parameter slope and the probability of insignificance of the estimated coefficient of the verification date meet the preset formula, if not, send a trend warning message for the verification device, where the preset formula is:

[0071] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0072] where pvalue represents the probability of insignificance of the estimated coefficient of the verification date, p acceptThe preset threshold value of the probability that the estimated coefficient representing the verification date is not significant, b accept Represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

[0073] In this embodiment, for the verification device error time series EquipError ts , the least squares fitting of the time series is performed using the Im function in R language to obtain the slope b and intercept intercept of the fitting parameter, and the probability pvalue that the estimated coefficient of the verification date in the time series is not significant. Then, when judging whether the fitting parameter slope and the probability that the estimated coefficient of the verification date is not significant satisfy the verification device error time series trend warning formula, if not satisfied, a verification device trend warning message is issued. Among them, the verification device error time series trend warning formula is:

[0074] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0075] Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, and p accept represents the preset threshold value of the probability that the estimated coefficient of the verification date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

[0076] The verification device error time series trend warning formula can be understood as: when the probability pvalue that the estimated coefficient of the verification date is not significant satisfies the hypothesis, and the slope of the fitting result exceeds the allowable slope value, a verification device trend warning is issued to remind the staff that the verification device error value changes significantly linearly with the verification date to take necessary management measures to prevent the costs and risks brought by the over - tolerance of the verification error.

[0077] tr represents the warning type, and the value is 0: no warning, 1: negative slope linear relationship over - tolerance warning, 2: positive slope linear relationship over - tolerance warning.

[0078] It should be noted that the preset value of the probability that the estimated coefficient is not significant is p accept . Then the original hypothesis of the test is that the verification date coefficient is significantly 0. If pvalue < p accept , the original hypothesis is rejected, that is, the corresponding verification date coefficient is significantly not 0, indicating that the verification device error has a linear change with the verification date.

[0079] In the preferred embodiment, after intercepting the time series with a sliding window, a second result is obtained, and it is judged whether the second result meets the preset conditions. If not satisfied, a verification device repeatability warning message is issued, specifically:

[0080] Intercept the time series using a sliding window to obtain multiple sliding window sequences;

[0081] Calculate the mean and standard deviation of the error measurement values within each sliding window to obtain the mean and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds a preset value, a repeatability warning message for the calibration device is issued.

[0082] In a preferred embodiment, after intercepting the time series with a sliding window, a third result is obtained. Determine whether the third result meets the preset conditions. If not, a repeatability warning message for the calibration device is issued. Specifically:

[0083] Calculate the mean and standard deviation of the error measurement values within each sliding window to obtain the mean and standard deviation of each sliding window;

[0084] After constructing a standard deviation time series using the standard deviations of the respective sliding windows, perform a least squares fitting calculation on the standard deviation time series to obtain the fitting parameter slope, intercept of the standard deviation, and the probability of non-significance of the estimation coefficient of the calibration date. If the fitting parameter slope and the probability of non-significance of the estimation coefficient of the calibration date do not satisfy the preset formula, a trend warning message for the repeatability measurement result of the calibration device is issued. Among them, the preset formula is:

[0085] tr=(p value <p accept )*((b>0)+1)*(abs(b)>b accept )

[0086] Where pvalue represents the probability of non-significance of the estimation coefficient of the calibration date, p accept represents the preset threshold of the probability of non-significance of the estimation coefficient of the calibration date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0087] In this embodiment, a sliding window with a window width of 10 and a sliding step of 1 is used to intercept the error time series EquipError of the calibration device ts to obtain an interception result, and the interception result is the error measurement values of the calibration device included in each window.

[0088] Calculate the mean x window and standard deviation s window of the error measurement values of the calibration device within each sliding window. When the standard deviation value within the sliding window exceeds the experimental standard deviation limit s, that is, the repeatability assessment of the calibration device does not meet the requirements, a repeatability warning message for the calibration device is issued.

[0089] It should be noted that the experimental standard deviation limit is preferably the experimental standard deviation limit s allowed by the device specified in the "JJG 597 Verification Device Verification Regulation".

[0090] Then, the repeatability of the verification device is judged. First, calculate the mean x of the error measurement values of the verification device within each sliding window window and the standard deviation s window After that, multiple measurement standard deviations can be obtained, which are respectively Then, use each standard deviation to establish a time series, and then use the Im function in R language to perform least squares fitting of the time series to obtain the fitting parameters slope b, intercept intercept, and the probability pvalue that the estimated coefficient of the verification date in the time series is not significant. Then, when judging whether the fitting parameter slope and the probability that the estimated coefficient of the verification date are not significant meet the verification device error time series trend warning formula, if not, a verification device trend warning message is issued. Among them, the verification device error time series trend warning formula is:

[0091] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0092] where pvalue represents the probability that the estimated coefficient of the verification date is not significant, p accept represents the preset threshold of the probability that the estimated coefficient of the verification date is not significant, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0093] The verification device error time series trend warning formula can be understood as: when the probability pvalue that the estimated coefficient of the verification date is not significant meets the hypothesis, and the slope of the fitting result exceeds the allowed slope value, a verification device trend warning is issued to remind the staff that the verification device error value changes significantly linearly with the verification date, so as to take necessary management measures to prevent the costs and risks brought by the exceeding of verification errors.

[0094] tr represents the warning type, and the value is 0: no warning, 1: negative slope linear relationship out-of-tolerance warning, 2: positive slope linear relationship out-of-tolerance warning.

[0095] It should be noted that the preset value of the probability that the estimated coefficient is not significant is p accept . Then the original hypothesis of the test is that the verification date coefficient is significantly 0. If pvalue < p accept , the original hypothesis is rejected, that is, the corresponding verification date coefficient is significantly not 0, indicating that the verification device error has a linear change with the verification date.

[0096] In a preferred embodiment, after calculating the stability of the time series, a fourth result and a fifth result are obtained, and it is determined whether the fourth result and the fifth result meet the preset conditions. If they do not meet, a warning message for the trend of the repetitive measurement results of the verification device is issued. Specifically:

[0097] Calculate the mean value and standard deviation of the error measurement values in each sliding window to obtain the mean value and standard deviation of each sliding window;

[0098] Perform a first-order difference calculation using the mean value of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than the preset first-order difference preset value, a warning message for the stability of the verification device is issued;

[0099] Construct a difference time series using the first-order difference results of each sliding window, and perform a least squares fitting estimation on the difference time series to obtain the fitting parameter slope, intercept of the first-order difference, and the improbability of the significance of the estimation coefficient of the verification date. If the fitting parameter slope and the improbability of the significance of the estimation coefficient of the verification date do not meet the preset formula, a warning message for the trend of the stability measurement results of the verification device is issued. Among them, the preset formula is:

[0100] tr=(pvalue<p accept )*((b>0)+1)*(abs(b)>b accept )

[0101] Among them, pvalue represents the improbability of the significance of the estimation coefficient of the verification date, and p accept represents the preset threshold of the improbability of the significance of the estimation coefficient of the verification date, and b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0102] In this embodiment, after calculating the mean value x window and the standard deviation s window of the error measurement values of the verification device in each sliding window, the mean value of the error measurement values of each verification device obtained Calculate the first-order difference result of the error measurement values of the verification device according to the mean value, that is, subtract the previous value from the next value of the measurement mean value respectively to obtain If each first-order difference result value exceeds the preset value, it means that the stability assessment of the verification device does not meet the requirements, and a warning message for the stability of the verification device is issued.

[0103] It should be noted that the preset value is the value of the deterioration of the stability of the verification device specified in the "Verification Regulation of JJG597 Verification Device".

[0104] A time series is constructed based on the mean of multiple error measurements of the calibration device. The time series is then fitted using the least squares method using the Im function in the R language to obtain the fitting parameters slope b and intercept intercept, as well as the probability pvalue of the estimated coefficient of the calibration date in the time series being insignificant. It is then determined whether the fitting parameters slope and the probability of the estimated coefficient of the calibration date being insignificant satisfy the calibration device error time series trend warning formula. If not, a calibration device trend warning message is issued. The calibration device error time series trend warning formula is:

[0105] tr=(pvalue<p accept )*((b>0)+1)*(abs(b)>b accept )

[0106] Among them, pvalue represents the possibility that the estimated coefficient of the test date is not significant, p accept Indicates the preset threshold value of the possibility that the estimated coefficient of the test date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the slope of the preset fitting parameter.

[0107] The error time series trend warning formula of the calibration device indicates that the probability that the estimated coefficient of the calibration date in the first-order difference result of the error mean is not significant pvalue satisfies the assumption, and the slope value b allowed for the first-order difference result of the error mean to change with time is: accept If the preset value is exceeded, an out-of-tolerance warning will be issued for the trend of the stability measurement results of the calibration device.

[0108] This method uses an electric energy meter error distribution function model based on historical calibration data from the same batch of smart electric energy meters to achieve online real-time verification of multiple calibration devices, thereby increasing the verification frequency and increasing the number of calibration device verification results, facilitating further analysis and mining of the verification results. A time series is then constructed based on the verification results, and the time series is fitted using the least squares method. In particular, the probability of insignificant characteristic estimation coefficients is used to determine whether there is a linear relationship between the verification results and the date. Time series trend warning formulas are then used to issue alarms for the trends of calibration device errors, stability, and repeatability. When the repeatability and stability of the calibration device change significantly over time, timely warnings can be issued, improving detection efficiency and accuracy.

[0109] Embodiment 2

[0110] Accordingly, see Figure 2 , Figure 2 The present invention provides a measurement verification device result analysis device. As shown in the figure, the measurement verification device result analysis device includes:

[0111] Verification module 201, configured to perform real-time verification on the calibration device by using the error distribution function model of the electricity meter, and obtain a verification result. The error distribution function model of the electricity meter is constructed by the central limit theorem and the Bayesian hierarchical model, and the verification result includes multiple error measurement values;

[0112] Time series construction module 202, configured to sort according to each error measurement value and the corresponding calibration date of each error measurement value to obtain a time series;

[0113] Judgment module 203, configured to perform least squares fitting calculation on the time series to obtain a first result, judge whether the first result meets a preset condition. If not, send a trend warning message for the calibration device. After performing sliding window truncation calculation on the time series, obtain a second result and a third result, judge whether the second result and the third result meet the preset condition. If not, send a repeatability warning message for the calibration device. After performing stability calculation on the time series, obtain a fourth result and a fifth result, judge whether the fourth result and the fifth result meet the preset condition. If not, send a stability warning message for the calibration device.

[0114] In a preferred embodiment, the judgment module 203 includes a calculation unit 2031 and a judgment unit 2032,

[0115] wherein, the calculation unit 2031 is configured to perform least squares fitting on the time series to obtain the fitting parameter slope, intercept, and the probability of insignificance of the estimated coefficient of the calibration date;

[0116] The judgment unit 2032 is configured to judge whether the fitting parameter slope and the probability of insignificance of the estimated coefficient of the calibration date meet a preset formula. If not, send a trend warning message for the calibration device. The preset formula is:

[0117] tr=(pvalue<p accept )*((b>0)+1)*(abs(b)>b accept )

[0118] wherein, pvalue represents the probability of insignificance of the estimated coefficient of the calibration date, p accept represents the preset threshold of the probability of insignificance of the estimated coefficient of the calibration date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0119] In a preferred embodiment, the judgment module 203 further includes a truncation unit 2033 and a standard deviation calculation unit 2034,

[0120] wherein, the truncation unit 2033 is configured to truncate the time series by using a sliding window to obtain multiple sliding window sequences;

[0121] The standard deviation calculation unit 2034 is used to calculate the mean value and the standard deviation of the error measurement values within each sliding window, obtaining the average value and the standard deviation of each sliding window. If the standard deviation of each sliding window exceeds the preset value, a repeatability warning message for the calibration device is issued.

[0122] In a preferred embodiment, after intercepting the time series with a sliding window, a third result is obtained, and it is determined whether the third result meets the preset conditions. If not, a repeatability warning message for the calibration device is issued. Specifically:

[0123] Calculate the mean value and the standard deviation of the error measurement values within each sliding window, obtaining the average value and the standard deviation of each sliding window;

[0124] After constructing a standard deviation time series using the standard deviations of each sliding window, perform a least squares fitting calculation on the standard deviation time series to obtain the fitting parameters of the standard deviation, namely the slope, the intercept, and the non-significance probability of the estimated coefficient of the calibration date. If the fitting parameter slope and the non-significance probability of the estimated coefficient of the calibration date do not meet the preset formula, a trend warning message for the repeatability measurement result of the calibration device is issued. Among them, the preset formula is:

[0125] tr=(pvalue<p accept )*((b>0)+1)*(abs(b)>b accept )

[0126] Among them, pvalue represents the non-significance probability of the estimated coefficient of the calibration date, and p accept represents the preset threshold of the non-significance probability of the estimated coefficient of the calibration date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope.

[0127] In a preferred embodiment, after performing a stability calculation on the time series, a fourth result and a fifth result are obtained, and it is determined whether the fourth result and the fifth result meet the preset conditions. If not, a trend warning message for the repeatability measurement result of the calibration device is issued. Specifically:

[0128] Calculate the mean value and the standard deviation of the error measurement values within each sliding window, obtaining the average value and the standard deviation of each sliding window;

[0129] Perform a first-order difference calculation using the average values of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than the preset first-order difference preset value, a stability warning message for the calibration device is issued;

[0130] Construct a difference time series using the first-order difference results of each sliding window, and perform least squares fitting estimation on the difference time series to obtain the fitting parameter slope, intercept of the first-order difference, and the probability of insignificance of the estimated coefficient of the verification date. If the fitting parameter slope and the probability of insignificance of the estimated coefficient of the verification date do not satisfy the preset formula, a trend warning message for the measurement result of the stability of the verification device is issued. The preset formula is:

[0131] tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept )

[0132] where pvalue represents the probability of insignificance of the estimated coefficient of the verification date, and p accept represents the preset threshold of the probability of insignificance of the estimated coefficient of the verification date, b accept represents the fitting parameter slope, and abs(b) represents the preset fitting parameter slope. The more detailed working principle and step process of this embodiment can but are not limited to refer to the relevant records of Embodiment 1.

[0133] In summary, implementing the embodiments of the present invention has the following beneficial effects:

[0134] Use the electric energy meter error distribution function model to perform real-time verification on the verification device to obtain verification results. The electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model. The verification results include multiple error measurement values. Organize according to each error measurement value and the corresponding verification date to obtain a time series. Perform least squares fitting calculation on the time series to obtain a first result, and judge whether the first result satisfies the preset conditions. If not, issue a trend warning message for the verification device. After performing sliding window truncation calculation on the time series, obtain a second result and a third result, and judge whether the second result and the third result satisfy the preset conditions. If not, issue a repeatability warning message for the verification device. After performing stability calculation on the time series, obtain a fourth result and a fifth result, and judge whether the fourth result and the fifth result satisfy the preset conditions. If not, issue a stability warning message for the verification device. This method uses the electric energy meter error distribution function model to perform real-time verification on the verification device, and after obtaining the verification results, analyzes and judges the verification results to obtain warning messages, improving the verification efficiency and accuracy.

[0135] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only for the specific embodiments of the present invention and is not intended to limit the protection scope of the present invention. In particular, for those skilled in the art, any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for analyzing the results of a metrological verification device, characterized in that, Including: Using an error distribution function model of an electricity meter to perform real-time verification on a verification device to obtain verification results. Among them, the error distribution function model of the electricity meter is constructed through the central limit theorem and the Bayesian hierarchical model, and the verification results include multiple error measurement values; Sorting according to each of the error measurement values and the verification dates corresponding to each of the error measurement values to obtain a time series; Performing least squares fitting calculation on the time series to obtain a first result, and determining whether the first result meets a preset condition. If it does not meet, a trend warning message for the verification device is issued. After performing sliding window truncation calculation on the time series, a second result and a third result are obtained, and determining whether the second result and the third result meet the preset condition. If they do not meet, a repeatability warning message for the verification device is issued. After performing stability calculation on the time series, a fourth result and a fifth result are obtained, and determining whether the fourth result and the fifth result meet the preset condition. If they do not meet, a stability warning message for the verification device is issued; The performing least squares fitting calculation on the time series to obtain a first result, and determining whether the first result meets a preset condition. If it does not meet, a trend warning message for the verification device is issued, specifically: Performing least squares fitting on the time series to obtain the fitting parameter slope, intercept, and the non-significance probability of the estimation coefficient of the verification date; When determining whether the fitting parameter slope and the non-significance probability of the estimation coefficient of the verification date meet a preset formula, if they do not meet, a trend warning message for the verification device is issued, where the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, p accept represents the preset threshold value of the probability that the estimated coefficient of the verification date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

2. The result analysis method of a metrological verification device according to claim 1, characterized in that After performing sliding window truncation on the time series to obtain a second result, and determining whether the second result meets the preset condition. If it does not meet, a repeatability warning message for the verification device is issued, specifically: Using a sliding window to truncate the time series to obtain multiple sliding window sequences; Performing mean value calculation and standard deviation calculation on the error measurement values within each sliding window to obtain the average value and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds a preset value, a repeatability warning message for the verification device is issued.

3. The result analysis method of a metrological verification device according to claim 1, characterized in that After performing sliding window truncation on the time series to obtain a third result, and determining whether the third result meets the preset condition. If it does not meet, a repeatability warning message for the verification device is issued, specifically: Performing mean value calculation and standard deviation calculation on the error measurement values within each sliding window to obtain the average value and standard deviation of each sliding window; After constructing a standard deviation time series using the standard deviations of each sliding window, performing least squares fitting calculation on the standard deviation time series to obtain the fitting parameter slope, intercept, and the non-significance probability of the estimation coefficient of the verification date of the standard deviation. If the fitting parameter slope and the non-significance probability of the estimation coefficient of the verification date do not meet the preset formula, a trend warning message for the repeatability measurement result of the verification device is issued, where the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, and p accept represents the preset threshold value of the probability that the estimated coefficient of the verification date is not significant, and b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

4. A result analysis method for a metrological verification device according to claim 1, characterized in that, After calculating the stability of the time series, the fourth result and the fifth result are obtained, and it is judged whether the fourth result and the fifth result meet the preset conditions. If they do not meet, a warning message about the trend of the repeated measurement results of the verification device is issued. Specifically: Calculate the mean and standard deviation of the error measurement values in each sliding window to obtain the mean and standard deviation of each sliding window; Perform a first-order difference calculation using the mean of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than the preset first-order difference preset value, a warning message about the stability of the verification device is issued; Construct a difference time series using the first-order difference results of each sliding window, and perform a least-squares fitting estimation on the difference time series to obtain the fitting parameter slope, intercept, and the probability of insignificance of the estimation coefficient of the verification date. If the fitting parameter slope and the probability of insignificance of the estimation coefficient of the verification date do not meet the preset formula, a warning message about the trend of the stability measurement results of the verification device is issued. Among them, the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, p accept represents the preset threshold of the probability that the estimated coefficient of the verification date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

5. A result analysis device for a metrological verification device, characterized in that, Including: A verification module for performing real-time verification on the verification device using the electric energy meter error distribution function model to obtain a verification result. Among them, the electric energy meter error distribution function model is constructed through the central limit theorem and the Bayesian hierarchical model, and the verification result includes multiple error measurement values; A time series construction module for sorting according to each of the error measurement values and the verification dates corresponding to each of the error measurement values to obtain a time series; A judgment module for performing a least-squares fitting calculation on the time series to obtain a first result, judging whether the first result meets the preset conditions. If it does not meet, a warning message about the trend of the verification device is issued. After performing a sliding window truncation calculation on the time series, the second result and the third result are obtained, and it is judged whether the second result and the third result meet the preset conditions. If they do not meet, a warning message about the repeatability of the verification device is issued. After performing a stability calculation on the time series, the fourth result and the fifth result are obtained, and it is judged whether the fourth result and the fifth result meet the preset conditions. If they do not meet, a warning message about the stability of the verification device is issued; The judgment module includes a calculation unit and a judgment unit, wherein the calculation unit is used to perform a least-squares fitting on the time series to obtain the fitting parameter slope, intercept, and the probability of insignificance of the estimation coefficient of the verification date; The judgment unit is used to judge whether the fitting parameter slope and the probability of insignificance of the estimation coefficient of the verification date meet the preset formula. If they do not meet, a warning message about the trend of the verification device is issued. Among them, the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, and p accept represents the preset threshold of the probability that the estimated coefficient of the verification date is not significant, and b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

6. The result analysis device of a metrological verification device according to claim 5, characterized in that, The judgment module further includes a truncation unit and a standard deviation calculation unit, wherein the truncation unit is used to truncate the time series using a sliding window to obtain a plurality of sliding window sequences; The standard deviation calculation unit is used to calculate the mean and standard deviation of the error measurement values within each sliding window, obtaining the mean and standard deviation of each sliding window. If the standard deviation of each sliding window exceeds the preset value, a repeatability warning message for the calibration device is issued.

7. The result analysis device of a metrological verification device according to claim 5, characterized in that, After intercepting the sliding window of the time series, a third result is obtained. It is determined whether the third result meets the preset conditions. If not, a repeatability warning message for the calibration device is issued. Specifically: Calculate the mean and standard deviation of the error measurement values within each sliding window, obtaining the mean and standard deviation of each sliding window. After constructing a standard deviation time series using the standard deviations of each sliding window, perform a least squares fitting calculation on the standard deviation time series to obtain the fitting parameter slope, intercept of the standard deviation, and the probability of insignificance of the estimation coefficient of the calibration date. If the fitting parameter slope and the probability of insignificance of the estimation coefficient of the calibration date do not satisfy the preset formula, a trend warning message for the repeatability measurement result of the calibration device is issued. Among them, the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, p accept represents the preset threshold value of the probability that the estimated coefficient of the verification date is not significant, b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

8. The result analysis device of a metrological verification device according to claim 5, characterized in that, After performing stability calculation on the time series, a fourth result and a fifth result are obtained. It is determined whether the fourth result and the fifth result meet the preset conditions. If not, a trend warning message for the repeatability measurement result of the calibration device is issued. Specifically: Calculate the mean and standard deviation of the error measurement values within each sliding window, obtaining the mean and standard deviation of each sliding window. Perform a first-order difference calculation using the means of each sliding window to obtain the first-order difference results of each sliding window. If the first-order difference results of each sliding window are greater than the preset first-order difference preset value, a stability warning message for the calibration device is issued. Construct a difference time series using the first-order difference results of each sliding window, and perform a least squares fitting estimation on the difference time series to obtain the fitting parameter slope, intercept of the first-order difference, and the probability of insignificance of the estimation coefficient of the calibration date. If the fitting parameter slope and the probability of insignificance of the estimation coefficient of the calibration date do not satisfy the preset formula, a trend warning message for the stability measurement result of the calibration device is issued. Among them, the preset formula is: tr = (pvalue < p accept ) * ((b > 0) + 1) * (abs(b) > b accept ) Among them, pvalue represents the probability that the estimated coefficient of the verification date is not significant, and p accept represents the preset threshold value of the probability that the estimated coefficient of the verification date is not significant, and b accept represents the slope of the fitting parameter, and abs(b) represents the preset slope of the fitting parameter.

Images