A real-time detection method for operation status of electric meter box

Abstract

The present application relates to the technical field of electric meter detection, and specifically to a real-time detection method for the operation status of an electric meter box, the method comprising: collecting the output power and operating temperature of the electric meter at each collection moment, and obtaining the output power data sequence, the operating temperature data sequence, each subsequence of the operating temperature data sequence and the output power data sequence of the electric meter, and the maximum value sequence and minimum value sequence of each subsequence of the output power data sequence; based on the distribution characteristics of the data points in the output power data sequence and the operating temperature data sequence, constructing the instantaneous fluctuation stability of each subsequence of the output power data sequence, and the fault anomaly confidence of each data point in each subsequence; and performing fault monitoring on the operation status of the electric meter. The present application aims to improve the accuracy of fault monitoring of the operation status of the electric meter using the isolation forest algorithm.

Description

Technical Field

[0001] The present application relates to the technical field of electric meter detection, and in particular to a method for real-time detection of the operating status of an electric meter box. Background Art

[0002] The electric meter box is a device used to measure, monitor and protect electric energy. It is usually composed of an electric meter, a power switch, a leakage protector, a relay, etc. These electric devices are subjected to electrical, thermal, mechanical and other effects and influences during operation. Long-term operation will cause aging and fatigue, resulting in a gradual decrease in the working efficiency of the equipment. Once the insulating medium in the electric meter box is damaged, it usually cannot be restored automatically, which seriously affects the operation of the electric equipment. Therefore, it is necessary to monitor the electric equipment in the electric meter box to ensure its stable operation, which is of great significance to the safety of the electric equipment in the electric meter box and the reliability of the operation of the power system. Monitoring the changes in the output power of the electric meter during operation is an effective fault monitoring method that can promptly detect and deal with potential problems of the electric meter.

[0003] The isolation forest algorithm is a commonly used abnormal data detection algorithm. It has good adaptability to various types of data and can quickly process large-scale data sets. It has great advantages in real-time monitoring of abnormal operation status of the meter box. However, the output power of the meter in the meter box will be affected by the operating temperature of the power equipment in the meter box and the ambient temperature, resulting in normal data fluctuations. For example, excessive temperature will cause the resistance of the power equipment in the meter box to change, thereby affecting the power size and stability, making it easy for the isolation forest algorithm to identify the normal data fluctuations of the meter during normal operation as abnormal data fluctuations generated when the meter fails. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a method for real-time detection of the operating status of an electric meter box. The technical solutions adopted are as follows:

[0005] The present invention provides a method for real-time detection of the operating status of an electric meter box, the method comprising the following steps:

[0006] Collect the output power and operating temperature of the electric meter in the electric meter box at each collection time;

[0007] Implementing fault monitoring of the operation status of the electric meter, the fault monitoring comprises the following steps:

[0008] (1) According to the output power and working temperature at all acquisition moments, the output power data sequence and the working temperature data sequence of the electric meter are obtained, and each subsequence of the working temperature data sequence and the output power data sequence is obtained, as well as the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence;

[0009] (2) Based on the differences in data points in the maximum value sequence and the minimum value sequence and the differences in the data point collection times, the instantaneous fluctuation stationarity of each subsequence of the output power data sequence is constructed;

[0010] (3) constructing the fault anomaly confidence level of each data point in each subsequence of the output power data sequence according to the distribution of data points in the subsequences of the output power data sequence and the operating temperature data sequence, and the instantaneous fluctuation smoothness of each subsequence;

[0011] (4) Performing fault monitoring on the operating status of the electric meter based on the output power data sequence, the operating temperature data sequence and the fault anomaly confidence level of each data point.

[0012] In one embodiment, the method for obtaining the output power data sequence, the operating temperature data sequence, each subsequence of the operating temperature data sequence, each subsequence of the output power data sequence, and the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence of the electric meter is:

[0013] The normalized values ​​of the output power at all the acquisition moments are combined into an output power data sequence of the electric meter in the order of the acquisition moments, and the calculation method of the output power data sequence is adopted for the working temperature to obtain the working temperature data sequence of the electric meter;

[0014] A mutation point detection algorithm is used to obtain mutation points in the output power data sequence, and the output power data sequence is divided into subsequences according to the mutation points. The minimum acquisition time and the maximum acquisition time of the data points in each subsequence are counted, and the operating temperature data sequence is intercepted according to each of the minimum acquisition time and the corresponding maximum acquisition time to obtain each subsequence;

[0015] The sequence decomposition method is used to obtain the trend item data sequence, maximum value sequence and minimum value sequence of each subsequence.

[0016] In one embodiment, the method of using the sequence decomposition method to respectively obtain the trend item data sequence, the maximum value sequence and the minimum value sequence of each subsequence is:

[0017] A time series decomposition method is used to respectively obtain a residual term data sequence of each subsequence of the output power data sequence and a trend term data sequence of each subsequence of the operating temperature data sequence;

[0018] The fitting curves of the residual data sequences of each subsequence of the output power data sequence are obtained, and all the maximum points and minimum points of each fitting curve are respectively arranged in the order of acquisition time to form the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence.

[0019] In one embodiment, the method for constructing the instantaneous fluctuation stability is:

[0020] The fluctuation stability of each subsequence of the output power data sequence is constructed based on the difference between the data points in the maximum value sequence and the minimum value sequence;

[0021] The fluctuation duration stability of each subsequence of the output power data sequence is constructed according to the difference in the data point collection time in the maximum value sequence and the minimum value sequence;

[0022] The product of the fluctuation degree stability and the fluctuation duration stability of each subsequence of the output power data sequence is taken as the instantaneous fluctuation stability of each subsequence of the output power data sequence.

[0023] In one embodiment, the method for constructing the fluctuation degree stability is:

[0024] The absolute values ​​of the differences between the corresponding position data in the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence are calculated, and recorded as the absolute value of the degree of fluctuation. The difference between any two absolute values ​​of the degree of fluctuation corresponding to each subsequence of the output power data sequence is calculated, and recorded as the degree of fluctuation difference. The mean of the squares of all the degree of fluctuation differences corresponding to each subsequence of the output power data sequence is calculated, and recorded as the degree of fluctuation mean. The degree of fluctuation smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding degree of fluctuation mean.

[0025] In one embodiment, the method for constructing the fluctuation duration stability is:

[0026] Calculate the absolute value of the difference between the collection moments of the corresponding position data in the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence, which is recorded as the absolute value of the fluctuation duration. Calculate the difference between any two absolute values ​​of the fluctuation duration corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration difference. Calculate the mean of the squares of all the fluctuation duration differences corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration mean. The fluctuation duration smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding fluctuation duration mean.

[0027] In one embodiment, the method for constructing the fault anomaly confidence level is:

[0028] Constructing a temperature influence factor of each subsequence of the output power data sequence according to an adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to each subsequence of the output power data sequence;

[0029] According to the distribution of data points in adjacent subsequences of the output power data sequence and the operating temperature data sequence, as well as the temperature influence factor and instantaneous fluctuation stability of each subsequence of the output power data sequence, a fault influence index of each subsequence of the output power data sequence is constructed;

[0030] The mean of the absolute values ​​of the differences between each data point and all other data points in the residual item data sequence of each subsequence of the output power data sequence is calculated, and recorded as the comprehensive mean. The difference between the total number of data points in each residual item data sequence and the sequence number of each data point is calculated, and recorded as the posterior difference. The sequence number of each data point in each residual item data sequence and the minimum value corresponding to the posterior difference are counted. The fault anomaly confidence of each data point in each subsequence of the output power data sequence is positively correlated with the fault influence index of each subsequence of the output power data sequence, the comprehensive mean corresponding to each data point in each subsequence of the output power data sequence, and the minimum value corresponding to each data point in each subsequence of the output power data sequence.

[0031] In one embodiment, the method for constructing the temperature influence factor is:

[0032] For each subsequence of the output power data sequence, for the adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to the subsequence, a straight line fitting method is used to obtain a fitting straight line of the trend item data sequence of the adjacent previous subsequence, and the slope of the fitting straight line is used as the exponent of an exponential function with a natural constant as the base, and the mean of all data points in the adjacent previous subsequence is calculated and recorded as the temperature data mean. The temperature influence factor of the subsequence of the output power data sequence is positively correlated with the temperature data mean and the calculation result of the exponential function, respectively;

[0033] Among them, when calculating the temperature influence factor of the first subsequence of the output power data sequence, the temperature data mean of the adjacent previous subsequence is assigned to the temperature data mean of the first subsequence of the working temperature data sequence; and the slope of the fitting straight line of the trend item data sequence of the adjacent previous subsequence is assigned to 0.

[0034] In one embodiment, the method for constructing the fault impact index is:

[0035] The difference between the mean of all data points in each subsequence of the output power data sequence and the mean of all data points in the adjacent previous subsequence is used as the independent variable of the sign function, which is recorded as the first sign function. For each subsequence of the operating temperature data sequence and the adjacent previous subsequence, the calculation method of the first sign function is used to obtain a sign function, which is recorded as the second sign function. The calculation results of each first sign function and the corresponding second sign function are used as the independent variable of the equals function;

[0036] Calculate the product of the instantaneous fluctuation stability and the temperature influence factor of each subsequence of the output power data sequence, recorded as a mixed product, the fault impact index of each subsequence of the output power data sequence is negatively correlated with the mixed product of each subsequence of the output power data sequence, and is positively correlated with the calculation result of the corresponding equals function of each subsequence of the output power data sequence;

[0037] Wherein, when calculating the fault impact index of the first subsequence of the output power data sequence, the independent variables of the first sign function and the second sign function are both assigned a value of 0.

[0038] In one embodiment, the fault monitoring of the operation status of the electric meter includes:

[0039] The values ​​of the data points in the output power data sequence at the same collection time are x-coordinates, and the values ​​of the data points in the operating temperature data sequence are y-coordinates to construct a rectangular coordinate system, and an anomaly detection algorithm is used to obtain an anomaly score of each data point in the rectangular coordinate system;

[0040] The product of the normalized result of the fault anomaly confidence corresponding to each data point in the rectangular coordinate system and the anomaly score is used as the anomaly score correction value of each data point in the rectangular coordinate system;

[0041] A threshold segmentation algorithm is used to obtain an abnormal threshold of the abnormal score correction value of all data points in the rectangular coordinate system, and the number of data points whose abnormal score correction value is greater than the abnormal threshold is counted. When the ratio of the number to the number of all data points in the rectangular coordinate system is greater than a preset fault monitoring threshold, it indicates that the monitored electric meter has a fault; otherwise, it indicates that the monitored electric meter has no fault.

[0042] This application has at least the following beneficial effects:

[0043] The present application proposes a real-time detection method for the operation status of an electric meter box, constructs instantaneous fluctuation stability based on a maximum value sequence and a minimum value sequence constructed from an output power data sequence, and analyzes the proximity of the instantaneous fluctuation degree between the data points where instantaneous fluctuation occurs. The beneficial effect is that the distinction between the instantaneous fluctuation stability of output power data of different subsequences of the output power data sequence when instantaneous fluctuation occurs is improved;

[0044] Based on the subsequences of the operating temperature data sequence and the trend item data sequence, the fault anomaly confidence is constructed in combination with the instantaneous fluctuation smoothness, which fully considers the influence of the operating temperature on the output power fluctuation. The beneficial effect is that the probability of normal data points with abnormal fluctuations in the output power data sequence caused by the operating temperature being identified as data points with abnormal fluctuations in the output power data sequence caused by the fault of the electric meter is reduced;

[0045] The anomaly score in the isolation forest algorithm is adaptively corrected based on the fault anomaly confidence, which improves the accuracy of fault monitoring of the operating status of the electric meter using the isolation forest algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 A flowchart of a method for real-time detection of the operating status of an electric meter box provided in one embodiment of the present application;

[0048] Figure 2 It is a data distribution diagram of the output power data sequence;

[0049] Figure 3 The figure is a flowchart for obtaining the confidence level of fault anomalies. DETAILED DESCRIPTION

[0050] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a real-time detection method of the operating status of an electric meter box proposed in accordance with the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0051] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0052] The following is a detailed description of a method for real-time detection of the operating status of an electric meter box provided by the present application in conjunction with the accompanying drawings.

[0053] See also Figure 1 , which shows a flowchart of a method for real-time detection of the operating status of an electric meter box provided by an embodiment of the present application, the method comprising the following steps:

[0054] Step 1: Collect the output power and the working temperature of the electric meter in the electric meter box at each collection time.

[0055] The embodiment of the present application aims to realize dynamic monitoring of the operating status of the electric meters in the electric meter box by analyzing the operating data of the electric meters related to the operating status of the electric meter box.

[0056] In this embodiment, the electric meter is placed in the electric meter box to make it work normally, the output end of the electric meter is connected to the current clamp of the power analyzer, and the power analyzer is used to collect the output power of the electric meter in the electric meter box. The electric meter outputs the power and uses the temperature sensor to collect the temperature of the electric meter inside the electric meter box as the working temperature of the electric meter in the electric meter box. In order to eliminate the influence of the data dimension on the subsequent operation status monitoring, the output power and the working temperature are respectively subjected to the maximum-minimum normalization processing. Among them, the maximum-minimum normalization processing is a common data processing technology.

[0057] Number of data points collected by the power analyzer and temperature sensor and the time interval for data collection The implementer can set the value of , The values ​​are 3000 and 0.1s respectively.

[0058] Step 2, based on the output power and operating temperature at all acquisition times, implement fault monitoring of the operating status of the electric meter in the electric meter box, wherein the fault monitoring includes the following steps:

[0059] Step 2.1, according to the output power and working temperature at all acquisition moments, obtain the output power data sequence and the working temperature data sequence of the electric meter, obtain each subsequence of the working temperature data sequence and the output power data sequence, and the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence.

[0060] During the normal operation of the meter, its output power is constantly fluctuating, but these fluctuations are usually relatively stable. When the meter fails, such as when the internal circuit of the power equipment in the meter box is damaged, the current passing through the meter will be unstable, and the output power fluctuation will become unstable, causing abnormal data fluctuations in the output power at multiple times. On the other hand, it will also cause large fluctuations in the output power during the power reduction process, thereby destroying the stability of the output power.

[0061] Furthermore, the normalized values ​​of the output power and the operating temperature at all the acquisition moments are respectively combined into an output power data sequence and an operating temperature data sequence in the order of the acquisition moments.

[0062] Secondly, the output power data sequence is used as the input of the mutation point detection algorithm, and several mutation points in the output power data sequence are obtained. The number of mutation points is recorded as , then the output power data sequence The output power data sequence is divided into Segment data sequence, each segment of the obtained data sequence is recorded as a subsequence of the output power data sequence, each subsequence represents the power consumption data of the user side of the meter in different time periods, such as household appliances. For example, the behavior of the household appliances discharging at different discharge powers collected by the meter at different working temperatures, or the behavior of the meter failing within a certain period of time and causing unstable output power. The minimum collection time and the maximum collection time of the data points in each subsequence of the output power data sequence are counted, and each subsequence is obtained by intercepting the working temperature data sequence according to each minimum collection time and the corresponding maximum collection time. In one embodiment of the present application, M is 9. The data distribution diagram of the output power data sequence is shown in FIG. Figure 2 As shown, in the schematic diagram, there are 9 mutation points in the output power data sequence, and these 9 mutation points divide the output power data sequence into 10 segments of data, each of which corresponds to a subsequence of the output power data sequence.

[0063] In one embodiment of the present application, the BG (Bernaola Galvan) segmentation algorithm is used to detect mutation points in the output power data sequence. There are many existing mutation point detection algorithms, such as the Pettitt method, the Manner-Kendall (MK) mutation test method, etc. As other implementation methods, the implementer may also use other mutation point detection algorithms to detect mutation points in the output power data sequence, and the present application does not impose any special restrictions on this.

[0064] Further, each subsequence of the output power data sequence and each subsequence of the operating temperature data sequence are respectively used as input, and the residual item data sequence of each subsequence of the output power data sequence and the trend item data sequence of each subsequence of the operating temperature data sequence are respectively obtained by using the time series decomposition algorithm. The data points in each residual item data sequence are curve-fitted using the curve fitting algorithm, with the acquisition time of the data point in each residual item data sequence and the value of the data point as the horizontal and vertical coordinates, respectively, to obtain the fitting curve of each residual item data sequence. All the maximum value points and minimum value points of each fitting curve are obtained, and the acquisition time corresponding to the maximum value point and the minimum value point is the acquisition time corresponding to the data point with large instantaneous fluctuation in each subsequence of the output power data sequence. All the maximum value points of each fitting curve are arranged in the order of the acquisition time to form the maximum value sequence of each subsequence of the output power data sequence, and the maximum value sequence acquisition method is adopted to obtain the minimum value sequence of each subsequence of the output power data sequence.

[0065] In one embodiment of the present application, the STL (Seasonal and Trend decomposition using Loess) decomposition algorithm is used to decompose each subsequence of the output power data sequence and each subsequence of the operating temperature data sequence to obtain a residual item data sequence of each subsequence of the output power data sequence and a trend item data sequence of each subsequence of the operating temperature data sequence. There are many existing time series decomposition algorithms. As other implementation methods, the implementer may also use other time series decomposition algorithms to obtain the residual item data sequence and the trend item data sequence, and the present application does not impose any special restrictions on this.

[0066] In one embodiment of the present application, a cubic B-spline curve fitting algorithm is used to obtain a fitting curve for the data points in each residual item data sequence. There are many existing curve fitting algorithms. As other implementation methods, the implementer may also use other curve fitting algorithms to obtain a fitting curve for the data points in each residual item data sequence. The present application does not impose any special restrictions on this.

[0067] Step 2.2, construct the instantaneous fluctuation stability of each subsequence of the output power data sequence based on the difference between the data points in the maximum value sequence and the minimum value sequence and the difference between the data point acquisition times.

[0068] The instantaneous fluctuation stability is constructed to characterize the stability of the data points in each subsequence of the output power data sequence when the instantaneous fluctuation occurs.

[0069] First, the fluctuation stability of each subsequence of the output power data sequence is constructed according to the difference between the data points in the maximum value sequence and the minimum value sequence. The specific process is as follows:

[0070] In the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence, the absolute value of the difference between the corresponding position data of the same order value is calculated, which is recorded as the absolute value of the fluctuation degree. It should be noted that if the elements in the maximum value sequence and the minimum value sequence are different, then under the premise of not affecting the original element order value, the operation of padding 0 is performed in the sequence with the shortest length so that the lengths of the two sequences are the same.

[0071] Furthermore, the difference between any two absolute values ​​of the fluctuation degrees corresponding to each subsequence of the output power data sequence is calculated and recorded as the fluctuation degree difference. The mean of the squares of all the fluctuation degree differences corresponding to each subsequence of the output power data sequence is calculated and recorded as the fluctuation degree mean. The fluctuation degree smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding fluctuation degree mean.

[0072] It should be understood that negative correlation means that the changing trends of the dependent variable and the independent variable are in opposite directions, such as a subtraction relationship, a division relationship, etc. The implementer can set it according to the actual situation, and this application does not impose any special restrictions.

[0073] In one embodiment of the present application, the average value of the fluctuation degree corresponding to each subsequence of the output power data sequence is calculated. The preset parameter constant is greater than 0 The sum of the values ​​is recorded as the first parameter adjustment sum value, and the reciprocal of the first parameter adjustment sum value is used as the fluctuation degree stability of each subsequence of the output power data sequence. In this embodiment, the fluctuation degree stability The volatility of the mean The calculation formula is: ; Among them, the parameter constant The purpose is to avoid the denominator being 0. The value of the parameter constant is preset manually and can be set by the implementer. In this embodiment, the value of the parameter constant is 0.1.

[0074] It should be noted that: since the maximum points and the minimum points in the curve usually appear alternately, the maximum points and the minimum points at the corresponding positions of the same order values ​​in the maximum sequence and the minimum sequence of each subsequence of the output power data sequence are adjacent, and the absolute value of the fluctuation degree can characterize the instantaneous fluctuation degree of each data point in each subsequence of the output power data sequence when an instantaneous fluctuation occurs. The larger the value of the fluctuation degree smoothness, the closer the instantaneous fluctuation degrees between all the data points with instantaneous fluctuations in each subsequence of the output power data sequence are, and the smaller the possibility of drastic fluctuations in the data points in each subsequence of the output power data sequence.

[0075] Secondly, the fluctuation duration stability of each subsequence of the output power data sequence is constructed according to the difference in the data point collection time in the maximum value sequence and the minimum value sequence. The specific process is as follows:

[0076] Calculate the absolute value of the difference between the collection moments of the position data corresponding to the same order value in the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence, which is recorded as the absolute value of the fluctuation duration. Calculate the difference between any two absolute values ​​of the fluctuation duration corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration difference. Calculate the mean of the squares of all the fluctuation duration differences corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration mean. The fluctuation duration smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding fluctuation duration mean.

[0077] In one embodiment of the present application, the average value of the fluctuation duration corresponding to each subsequence of the output power data sequence is calculated. The preset parameter constant is greater than 0 The sum of the values ​​is recorded as the second parameter adjustment sum value, and the reciprocal of the second parameter adjustment sum value is used as the fluctuation duration stability of each subsequence of the output power data sequence. In this embodiment, the fluctuation duration stability The average fluctuation duration The calculation formula is: ; Among them, the parameter constant The purpose is to avoid the denominator being 0. The value of the parameter constant is preset manually and can be set by the implementer. In this embodiment, the value of the parameter constant is 0.1.

[0078] It should be noted that the absolute value of the fluctuation duration can represent the duration of the instantaneous fluctuation when the instantaneous fluctuation occurs in each data point in each subsequence of the output power data sequence. The larger the value of the fluctuation duration smoothness, the closer the duration of the instantaneous fluctuation between all the data points with instantaneous fluctuations in each subsequence of the output power data sequence, and the smaller the possibility of irregular fluctuations in the data points in each subsequence of the output power data sequence.

[0079] Finally, according to the fluctuation degree stability and fluctuation duration stability of each subsequence of the output power data sequence, the instantaneous fluctuation stability of each subsequence of the output power data sequence is constructed. The specific process is as follows:

[0080] The product of the fluctuation degree stability and the fluctuation duration stability of each subsequence of the output power data sequence is used as the instantaneous fluctuation stability of each subsequence of the output power data sequence. In this embodiment, the calculation formula of the instantaneous fluctuation stability C of each subsequence is: .

[0081] It should be noted that: the smaller the value of the instantaneous fluctuation smoothness, the more likely it is that the data points in each subsequence of the output power data sequence will experience irregular and violent fluctuations; the worse the smoothness of the data points in each subsequence of the output power data sequence when instantaneous fluctuations occur, the greater the possibility of abnormally fluctuating data points appearing in each subsequence of the output power data sequence.

[0082] At this point, the instantaneous fluctuation stability of each subsequence of the output power data sequence is obtained.

[0083] Step 2.3, construct the fault anomaly confidence of each data point in each subsequence of the output power data sequence according to the distribution of data points in the subsequences of the output power data sequence and the operating temperature data sequence, and the instantaneous fluctuation smoothness of each subsequence of the output power data sequence.

[0084] The fault anomaly confidence is constructed to characterize the possibility of abnormal fluctuations caused by the meter operation failure in the meter box at each data point in each subsequence of the output power data sequence. The schematic diagram of the acquisition process of the fault anomaly confidence is shown in Figure 3 shown.

[0085] First, according to the adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to each subsequence of the output power data sequence, the temperature influence factor of each subsequence of the output power data sequence is constructed. The specific process is as follows:

[0086] For each subsequence of the output power data sequence, for the adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to the subsequence, a straight line fitting method is used to obtain a fitting straight line of the trend item data sequence of the adjacent previous subsequence, and the slope of the fitting straight line is used as the exponent of an exponential function with a natural constant as the base, and the mean of all data points in the adjacent previous subsequence is calculated and recorded as the temperature data mean. The temperature influence factor of the subsequence of the output power data sequence is positively correlated with the temperature data mean and the calculation result of the exponential function, respectively.

[0087] Among them, when calculating the temperature influence factor of the first subsequence of the output power data sequence, the temperature data mean of the adjacent previous subsequence is assigned to the temperature data mean of the first subsequence of the working temperature data sequence; and the slope of the fitting straight line of the trend item data sequence of the adjacent previous subsequence is assigned to 0.

[0088] It should be understood that positive correlation means that the changing trends of the dependent variable and the independent variable are in the same direction, such as a multiplication relationship, an addition relationship, etc. Implementers can set it according to actual conditions.

[0089] In one embodiment of the present application, the calculation formula of the temperature influence factor is: ,in, is the temperature influence factor of the i-th subsequence of the output power data sequence, is an exponential function with a natural constant as base, is the slope of the fitted straight line corresponding to the adjacent previous subsequence of the operating temperature data sequence corresponding to the i-th subsequence, is the mean temperature data.

[0090] In one embodiment of the present application, the least squares method is used to obtain the fitting straight line for all data points in each trend item data sequence. As other implementation methods, the implementer may also use other straight line fitting algorithms to obtain the fitting straight line for all data points in each trend item data sequence. The present application does not impose any special restrictions on this.

[0091] It should be noted that: the larger the mean value of the temperature data of the adjacent previous subsequence is, and the larger the slope of the fitting straight line of the trend item data sequence of the adjacent previous subsequence is, the more the meter needs to reduce the output power in the time period of the corresponding output power data sequence subsequence to reduce the heat accumulation of the meter during operation, then the larger the value of the temperature influence factor is, the more the change of the data points in the output power data sequence subsequence is affected by the operating temperature of the meter, and the smaller the possibility that the abnormal data fluctuation in the output power data sequence subsequence is caused by a fault in the meter.

[0092] Secondly, according to the distribution of data points in adjacent subsequences of the output power data sequence and the operating temperature data sequence, as well as the temperature influence factor and instantaneous fluctuation stability of each subsequence of the output power data sequence, the fault influence index of each subsequence of the output power data sequence is constructed. The specific process is as follows:

[0093] The difference between the mean of all data points in each subsequence of the output power data sequence and the mean of all data points in the adjacent previous subsequence is used as the independent variable of the sign function to obtain the first sign function. For each subsequence of the operating temperature data sequence and the adjacent previous subsequence, the sign function is obtained by using the calculation method of the first sign function, which is recorded as the second sign function. The calculation results of each of the first sign function and the corresponding second sign function are used as the independent variable of the equals function;

[0094] It should be noted that the equals function is a function that determines whether two independent variables are equal. If the two independent variables are equal, the output of the equals function is 1; if the two independent variables are not equal, the output of the equals function is 0.

[0095] The product of the instantaneous fluctuation smoothness and the temperature influence factor of each subsequence of the output power data sequence is calculated and recorded as a mixed product. The fault impact index of each subsequence of the output power data sequence is negatively correlated with the mixed product of each subsequence of the output power data sequence, and is positively correlated with the calculation result of the corresponding equals function of each subsequence of the output power data sequence.

[0096] Wherein, when calculating the fault impact index of the first subsequence of the output power data sequence, the independent variables of the first sign function and the second sign function are both assigned a value of 0.

[0097] In one embodiment of the present application, the mixed product of each subsequence of the output power data sequence is calculated. The sum of the preset parameter constant greater than 0 is recorded as the third parameter sum value, and the calculation result of the equals function corresponding to each subsequence of the output power data sequence is The ratio of the third parameter adjustment value to the output power data sequence is used as the fault impact index of each subsequence of the output power data sequence. In this embodiment, the fault impact index of each subsequence is The calculation formula is: ; Among them, the parameter constant The purpose is to avoid the denominator being 0. The value of the parameter constant is preset manually and can be set by the implementer. In this embodiment, the value of the parameter constant is 0.1.

[0098] It should be noted that: Generally, in two adjacent time periods, if the temperature in the meter box shows an upward trend, in order to reduce the heat accumulation of the meter during operation, the output power of the meter will show a downward trend. Therefore, if the calculation results of the first sign function and the corresponding second sign function are the same, the calculation result of the equals function is 1, then the operating temperature and the output power change synchronously in two adjacent time periods, such as rising or falling at the same time, then the abnormal data fluctuations appearing in the subsequence of the output power data sequence are more likely to be caused by the meter failure; at the same time, the smaller the value of the instantaneous fluctuation smoothness of the subsequence of the output power data sequence, the more likely it is that abnormal data fluctuations appear in the subsequence of the output power data sequence; then the larger the value of the fault impact index, the more likely it is that abnormal data fluctuations appear in the subsequence of the output power data sequence caused by the meter failure.

[0099] Finally, according to the fault influence index of each subsequence of the output power data sequence and the residual term data sequence, the fault anomaly confidence of each data point in each subsequence of the output power data sequence is constructed. The specific process is as follows:

[0100] The mean of the absolute values ​​of the differences between each data point and all other data points in the residual item data sequence of each subsequence of the output power data sequence is calculated, and recorded as the comprehensive mean. The difference between the total number of data points and the order value of each data point in each residual item data sequence is calculated, and recorded as the posterior difference. The minimum value between the order value of each data point in each residual item data sequence and the posterior difference of each data point is obtained. The fault anomaly confidence of each data point in each subsequence of the output power data sequence is positively correlated with the fault influence index of each subsequence of the output power data sequence, the comprehensive mean corresponding to each data point in each subsequence of the output power data sequence, and the minimum value corresponding to each data point.

[0101] In one embodiment of the present application, the fault anomaly confidence level of each data point in each subsequence of the output power data sequence is The acquisition method is as follows: the fault impact index of each subsequence of the output power data sequence , each data point in each subsequence of the output power data sequence corresponds to the comprehensive mean value The product of the three corresponding to the minimum value x, that is .

[0102] It should be noted that: in the residual data sequence of each subsequence of the output power data sequence, the larger the mean of the absolute value of the difference between each data point and all other data points, the greater the fluctuation difference between each data point and all other data points, and the more likely each data point is a data point where abnormal fluctuation occurs in the corresponding output power data sequence subsequence; the closer the position of each data point in the output power data sequence subsequence is to the endpoint, the smaller the minimum value is, the greater the possibility that the abnormal fluctuation occurring in each data point is caused by a change in output power when the operating temperature changes, and the smaller the possibility that it is caused by a fault in the electric meter; then the larger the value of the fault anomaly confidence, the greater the possibility that the abnormal fluctuation occurring in each data point is caused by a fault in the electric meter.

[0103] At this point, the fault anomaly confidence level of each data point in each subsequence of the output power data sequence is obtained.

[0104] Step 2.4, performing fault monitoring on the operation status of the electric meter according to the output power data sequence, the operating temperature data sequence and the fault anomaly confidence level of each data point.

[0105] A rectangular coordinate system is constructed by using the values ​​of the data points in the output power data sequence at the same collection time as the x-coordinate and the values ​​of the data points in the operating temperature data sequence as the y-coordinate, and using the data points in the rectangular coordinate system as input of the isolation forest algorithm, and outputting the abnormality score of each data point in the rectangular coordinate system. The isolation forest algorithm is a well-known technology, and the specific process will not be repeated here.

[0106] It should be noted that the number of data points extracted each time in the isolation forest algorithm and the number of isolated trees in the forest The values ​​are 256 and 100 respectively. and The implementer can also set the value of by himself, and this application does not make any special limitation.

[0107] Furthermore, the fault anomaly confidence corresponding to each data point in the rectangular coordinate system is normalized by maximum and minimum standardization, and the product of the normalized result corresponding to each data point and the anomaly score is used as the anomaly score correction value of each data point in the rectangular coordinate system.

[0108] It should be noted that: the larger the fault anomaly confidence value of the data point, the more likely that the anomaly of the data point is caused by a fault in the electric meter, the larger the anomaly score value of the data point in the isolation forest algorithm should be, the larger the anomaly score correction value of the data point should be, and the greater the possibility that the electric meter is faulty at the collection time corresponding to the data point.

[0109] Calculate the anomaly score correction value of each data point in the rectangular coordinate system, use all the obtained anomaly score correction values ​​as the input of the threshold segmentation algorithm, and output the anomaly threshold. Count the number of all data points in the rectangular coordinate system whose anomaly score correction value is greater than the anomaly threshold, and use the ratio of the obtained number to the number of all data points in the rectangular coordinate system as the fault monitoring value of the meter. The larger the fault monitoring value of the monitored meter, the more likely it is that the meter has a fault.

[0110] In one embodiment of the present application, the maximum inter-class variance algorithm is used to obtain the abnormal threshold. As other implementation methods, the implementer may also use other threshold segmentation algorithms to obtain the abnormal threshold, and the present application does not impose any special restrictions on this.

[0111] Furthermore, a fault monitoring threshold is set. If the fault monitoring value of the monitored meter is greater than the fault monitoring threshold, it indicates that the monitored meter has a fault and an alarm is issued to remind. Otherwise, it indicates that the monitored meter has no fault.

[0112] In one embodiment of the present application, the value of the fault monitoring threshold is 0.8. The value of the fault monitoring threshold is preset manually and can be set by the implementer according to actual conditions. This application does not impose any special restrictions.

[0113] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0114] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0115] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for real-time detection of the operating status of an electric meter box, characterized in that: The method comprises the following steps: Collect the output power and operating temperature of the electric meter in the electric meter box at each collection time; Implementing fault monitoring of the operation status of the electric meter, the fault monitoring comprises the following steps: (1) According to the output power and working temperature at all acquisition moments, the output power data sequence and the working temperature data sequence of the electric meter are obtained, and each subsequence of the working temperature data sequence and the output power data sequence is obtained, as well as the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence; (2) Based on the differences in data points in the maximum value sequence and the minimum value sequence and the differences in the data point collection times, the instantaneous fluctuation stationarity of each subsequence of the output power data sequence is constructed; (3) constructing the fault anomaly confidence level of each data point in each subsequence of the output power data sequence according to the distribution of data points in the subsequences of the output power data sequence and the operating temperature data sequence, and the instantaneous fluctuation smoothness of each subsequence; (4) performing fault monitoring on the operating state of the electric meter according to the output power data sequence, the operating temperature data sequence and the fault anomaly confidence level of each data point; The method for obtaining the output power data sequence, the operating temperature data sequence, each subsequence of the operating temperature data sequence, each subsequence of the output power data sequence, and the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence of the electric meter is: The normalized values ​​of the output power at all the acquisition moments are combined into an output power data sequence of the electric meter in the order of the acquisition moments, and the calculation method of the output power data sequence is adopted for the working temperature to obtain the working temperature data sequence of the electric meter; A mutation point detection algorithm is used to obtain mutation points in the output power data sequence, and the output power data sequence is divided into subsequences according to the mutation points. The minimum acquisition time and the maximum acquisition time of the data points in each subsequence are counted, and the operating temperature data sequence is intercepted according to each of the minimum acquisition time and the corresponding maximum acquisition time to obtain each subsequence; The trend item data sequence, maximum value sequence and minimum value sequence of each subsequence are obtained by using the sequence decomposition method; The method of using the sequence decomposition method to obtain the trend item data sequence, the maximum value sequence and the minimum value sequence of each subsequence is as follows: A time series decomposition method is used to respectively obtain a residual term data sequence of each subsequence of the output power data sequence and a trend term data sequence of each subsequence of the operating temperature data sequence; The fitting curves of the residual data sequences of each subsequence of the output power data sequence are obtained, and all the maximum points and minimum points of each fitting curve are respectively arranged in the order of acquisition time to form the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence.

2. A method for real-time detection of the operating status of an electric meter box according to claim 1, characterized in that: The method for constructing the instantaneous fluctuation stability is: The fluctuation stability of each subsequence of the output power data sequence is constructed based on the difference between the data points in the maximum value sequence and the minimum value sequence; The fluctuation duration stability of each subsequence of the output power data sequence is constructed according to the difference in the data point collection time in the maximum value sequence and the minimum value sequence; The product of the fluctuation degree stability and the fluctuation duration stability of each subsequence of the output power data sequence is taken as the instantaneous fluctuation stability of each subsequence of the output power data sequence.

3. A method for real-time detection of the operating status of an electric meter box as claimed in claim 2, characterized in that: The method for constructing the fluctuation degree stability is: The absolute values ​​of the differences between the corresponding position data in the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence are calculated, and recorded as the absolute value of the degree of fluctuation. The difference between any two absolute values ​​of the degree of fluctuation corresponding to each subsequence of the output power data sequence is calculated, and recorded as the degree of fluctuation difference. The mean of the squares of all the degree of fluctuation differences corresponding to each subsequence of the output power data sequence is calculated, and recorded as the degree of fluctuation mean. The degree of fluctuation smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding degree of fluctuation mean.

4. A method for real-time detection of the operating status of an electric meter box as claimed in claim 2, characterized in that: The method for constructing the fluctuation duration stability is: Calculate the absolute value of the difference between the collection moments of the corresponding position data in the maximum value sequence and the minimum value sequence of each subsequence of the output power data sequence, which is recorded as the absolute value of the fluctuation duration. Calculate the difference between any two absolute values ​​of the fluctuation duration corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration difference. Calculate the mean of the squares of all the fluctuation duration differences corresponding to each subsequence of the output power data sequence, which is recorded as the fluctuation duration mean. The fluctuation duration smoothness of each subsequence of the output power data sequence is negatively correlated with the corresponding fluctuation duration mean.

5. A method for real-time detection of the operating status of an electric meter box as claimed in claim 1, characterized in that: The method for constructing the fault anomaly confidence level is as follows: Constructing a temperature influence factor of each subsequence of the output power data sequence according to an adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to each subsequence of the output power data sequence; According to the distribution of data points in adjacent subsequences of the output power data sequence and the operating temperature data sequence, as well as the temperature influence factor and instantaneous fluctuation stability of each subsequence of the output power data sequence, a fault influence index of each subsequence of the output power data sequence is constructed; The mean of the absolute values ​​of the differences between each data point and all other data points in the residual item data sequence of each subsequence of the output power data sequence is calculated, and recorded as the comprehensive mean. The difference between the total number of data points in each residual item data sequence and the sequence number of each data point is calculated, and recorded as the posterior difference. The sequence number of each data point in each residual item data sequence and the minimum value corresponding to the posterior difference are counted. The fault anomaly confidence of each data point in each subsequence of the output power data sequence is positively correlated with the fault influence index of each subsequence of the output power data sequence, the comprehensive mean corresponding to each data point in each subsequence of the output power data sequence, and the minimum value corresponding to each data point in each subsequence of the output power data sequence.

6. A method for real-time detection of the operating status of an electric meter box as claimed in claim 5, characterized in that: The method for constructing the temperature influence factor is: For each subsequence of the output power data sequence, for the adjacent previous subsequence of the operating temperature data sequence subsequence corresponding to the subsequence, a straight line fitting method is used to obtain a fitting straight line of the trend item data sequence of the adjacent previous subsequence, and the slope of the fitting straight line is used as the exponent of an exponential function with a natural constant as the base, and the mean of all data points in the adjacent previous subsequence is calculated and recorded as the temperature data mean. The temperature influence factor of the subsequence of the output power data sequence is positively correlated with the temperature data mean and the calculation result of the exponential function, respectively; Among them, when calculating the temperature influence factor of the first subsequence of the output power data sequence, the temperature data mean of the adjacent previous subsequence is assigned to the temperature data mean of the first subsequence of the working temperature data sequence; and the slope of the fitting straight line of the trend item data sequence of the adjacent previous subsequence is assigned to 0.

7. A method for real-time detection of the operating status of an electric meter box as claimed in claim 5, characterized in that: The method for constructing the fault impact index is: The difference between the mean of all data points in each subsequence of the output power data sequence and the mean of all data points in the adjacent previous subsequence is used as the independent variable of the sign function, which is recorded as the first sign function. For each subsequence of the operating temperature data sequence and the adjacent previous subsequence, the calculation method of the first sign function is used to obtain a sign function, which is recorded as the second sign function. The calculation results of each first sign function and the corresponding second sign function are used as the independent variable of the equals function; Calculate the product of the instantaneous fluctuation stability and the temperature influence factor of each subsequence of the output power data sequence, recorded as a mixed product, the fault impact index of each subsequence of the output power data sequence is negatively correlated with the mixed product of each subsequence of the output power data sequence, and is positively correlated with the calculation result of the corresponding equals function of each subsequence of the output power data sequence; Wherein, when calculating the fault impact index of the first subsequence of the output power data sequence, the independent variables of the first sign function and the second sign function are both assigned a value of 0.

8. A method for real-time detection of the operating status of an electric meter box as claimed in claim 1, characterized in that: The fault monitoring of the operation status of the electric meter includes: The values ​​of the data points in the output power data sequence at the same collection time are x-coordinates, and the values ​​of the data points in the operating temperature data sequence are y-coordinates to construct a rectangular coordinate system, and an anomaly detection algorithm is used to obtain an anomaly score of each data point in the rectangular coordinate system; The product of the normalized result of the fault anomaly confidence corresponding to each data point in the rectangular coordinate system and the anomaly score is used as the anomaly score correction value of each data point in the rectangular coordinate system; A threshold segmentation algorithm is used to obtain an abnormal threshold of the abnormal score correction value of all data points in the rectangular coordinate system, and the number of data points whose abnormal score correction value is greater than the abnormal threshold is counted. When the ratio of the number to the number of all data points in the rectangular coordinate system is greater than a preset fault monitoring threshold, it indicates that the monitored electric meter has a fault; otherwise, it indicates that the monitored electric meter has no fault.

Images