A method and device for predicting the remaining life of an intelligent electric energy meter
By adaptively acquiring scale parameters and combining the trend of current data with the degree of load change, the Weibull distribution model is adjusted, which solves the prediction bias problem caused by inappropriate selection of Weibull distribution scale parameters and realizes efficient and accurate prediction of the remaining life of smart energy meters.
Patent Information
- Application Number
- CN202411172664.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2044-08-26
AI Technical Summary
In existing technologies for predicting the lifespan of smart energy meters, the selection of the scale parameter of the Weibull distribution is inappropriate, which affects the prediction effect and leads to prediction bias. In addition, traditional detection methods consume a lot of manpower and resources and have a low degree of informatization.
By collecting current data, dividing it into multiple windows for fitting, obtaining trend feature values and similarity, adaptively obtaining scale parameters, and combining the trend of current data and the degree of load change, adjusting the Weibull distribution model to improve prediction accuracy.
It improves the accuracy of remaining life prediction for smart energy meters, avoids prediction deviations caused by unreasonable parameter settings, and achieves efficient and intelligent life assessment.
Smart Images

Figure CN118707433B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing technology, and more specifically, to a method and device for predicting the remaining life of a smart electric energy meter. Background Art
[0002] Since 2009, smart energy meters have been put into use on a large scale across the country. By June 2024, the cumulative installation of smart energy meters will exceed 612 million. However, the use of traditional means to conduct on-site inspections and random inspections of a large number of energy meters will consume a lot of manpower and material resources; at the same time, compared with the more advanced smart energy meters, regular on-site weekly inspections and random sampling inspections also have problems such as low degree of informatization, large amount of ineffective work, and poor timeliness of fault detection, resulting in a lot of waste. Therefore, it is necessary to use data information to carry out relevant research on the reliability life prediction method of smart energy meters; on the other hand, a large amount of data information has been accumulated in the information systems such as the power consumption information collection system of the State Grid Power Company and the metering production and dispatching platform, which also provides data support for efficient and intelligent operation status monitoring, reliability prediction and life evaluation of energy meters.
[0003] In the 2019 issue of the Journal of Electrical Measurement and Instrumentation, an article entitled "Research on Optimization of Weibull Distribution Based on Remaining Life Prediction of Electricity Meters" was published. The original author proposed a method combining genetic algorithms to obtain optimal parameters. The genetic algorithm was used to estimate the parameters of the maximum likelihood function of the Weibull distribution, determine the reliability function of the remaining life of a batch of electricity meters, and finally calculate the remaining life.
[0004] The above genetic algorithm has many advantages in parameter optimization and finding the global optimal solution in the search space, but it usually requires a large number of parameter combination generation, evaluation and selection operations, which consume a lot of computing resources and time on large-scale problems. In addition, the genetic algorithm also requires some parameters. The adjustment of these parameters has a great influence on the performance and effect of the algorithm. The scale parameter in the Weibull distribution determines the position and scale in the Weibull distribution. Therefore, the selection of the scale parameter will affect the final prediction effect of the model on the life of the smart electricity meter. Summary of the invention
[0005] In order to solve the technical problem that the scale parameter in the Weibull distribution determines the position and scale in the Weibull distribution, and the inappropriate selected scale parameter will affect the final prediction effect of the model on the life of the smart electric energy meter, the present invention proposes a method and device for predicting the remaining life of a smart electric energy meter.
[0006] In a first aspect, the present invention provides a method for predicting the remaining life of a smart electric energy meter, which adopts the following technical solution:
[0007] A method for predicting the remaining life of a smart electric energy meter comprises the following steps:
[0008] Collect current data; divide the current data into multiple windows; fit the current data in each window to obtain the fitting value of each current data in each window; obtain the trend characteristic value of each window based on the correlation between the fitting value and the actual value of each current data in each window; obtain the similarity between adjacent windows according to the difference of the trend characteristic values of adjacent windows; obtain the trend of the current data according to the similarity between adjacent windows;
[0009] Obtain multiple maximum values in the current data; obtain the load change degree corresponding to each maximum value in the current data according to the numerical change of the maximum value and the current data near it; and take the average of the absolute values of the differences in the load change degrees corresponding to all adjacent maximum values as the load change degree of the reference device;
[0010] Get scale parameters ; In the formula, represents the scale parameter; Represents the degree of load change of the reference device; Represents the trend of current data; exp() represents an exponential function with a natural constant as the base; represents the preset hyperparameters, Represents the base of the logarithmic function; the remaining life of the smart electric energy meter that has not failed is predicted according to the scale parameter.
[0011] The innovation of the present invention lies in that the scale parameters are adaptively acquired according to the trend of the current data, that is, the fluctuation or change trend of the current data, and the load change degree of the reference device to predict the remaining life of the smart electric energy meter that has not yet failed. The scale parameters can be adjusted according to the distribution characteristics of the collected current data, thereby improving the accuracy of the prediction results and avoiding the prediction deviation caused by unreasonable parameter setting of the Weibull distribution fitting model.
[0012] Preferably, dividing the current data into a plurality of windows comprises:
[0013] The number of window data N is preset, and according to the number of window data N, the current data sequence is divided into a plurality of windows in sequence, and the number of current data in each window is N.
[0014] Preferably, fitting the current data in each window to obtain the fitting value of each current data in each window includes:
[0015] The current data in each window are fitted with a regression line using the least squares method to obtain a trend line of the current data in each window and a fitting value of each current data in each window.
[0016] Preferably, the step of obtaining the trend characteristic value of each window includes:
[0017] Obtain the absolute value of the difference between the fitted value of the nth current data in the jth window and the fitted value of the n+1th current data as the first-order difference of the fitted value of the nth current data in the jth window, and obtain the first-order difference of the fitted value of each current data in each window;
[0018] Obtain the absolute value of the difference between the actual value of the nth current data in the jth window and the actual value of the n+1th current data as the first-order difference of the actual value of the nth current data in the jth window, and obtain the first-order difference of the actual value of each current data in each window;
[0019] The DTW algorithm is used to obtain the first-order difference of the fitted value of all current data in each window and the correlation coefficient of the first-order difference of the actual value as the trend characteristic value of each window.
[0020] The correlation between the first-order difference of the fitted value of all current data in the window and the actual first-order difference can reflect the trend change of the current data, and the similarity between adjacent windows can be obtained according to the trend characteristic value of each window.
[0021] Preferably, the obtaining of the similarity between adjacent windows includes:
[0022] ;
[0023] In the formula, Represents the similarity between the jth window and the j+1th window; Represents the trend characteristic value of the jth window; Represents the trend characteristic value of the j+1th window; represents the slope of the trend line of the current data in the jth window; The intercept of the trend line representing the current data in the jth window; The slope of the trend line representing the current data in the j+1th window; Represents the intercept of the trend line of the current data in the j+1th window; exp() represents an exponential function with a natural constant as the base.
[0024] Similarity represents whether the trends between adjacent windows are similar. If the trends between all adjacent windows are not similar, it means that the value of the current data fluctuates, that is, the greater the trend of the current data, the trend of the current data can be obtained based on the similarity.
[0025] Preferably, the obtaining of the trend of the current data includes:
[0026] ;
[0027] In the formula, represents the trend of current data; J represents the number of windows; Represents the similarity between the jth window and the j+1th window; Represents the similarity between the j+1th window and the j+2th window; || represents the absolute value symbol.
[0028] The greater the trend of the current data, the more it indicates that the value of the current data fluctuates, and a larger scale parameter is required at this time. Therefore, the scale parameter can be adaptively adjusted according to the trend of the current data.
[0029] Preferably, the step of obtaining the load variation degree corresponding to each maximum value in the current data includes:
[0030] ;
[0031] In the formula, Represents the load change degree corresponding to the pth maximum value in the current data; represents the value of the pth maximum; Represents the value of the d+1th maximum; represents the value of the p-1th maximum; Represents the value of the current data after the pth maximum value in the current data sequence; Represents the value of the previous current data of the pth maximum value in the current data sequence; || represents the absolute value sign.
[0032] This facilitates the subsequent acquisition of the load change degree of the reference device based on the load change degree corresponding to each maximum value in the current data.
[0033] In a second aspect, the present invention provides a device for predicting the remaining life of a smart electric energy meter, which adopts the following technical solution:
[0034] A device for predicting the remaining life of a smart energy meter comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for predicting the remaining life of a smart energy meter is implemented.
[0035] By adopting the above technical solution, the above-mentioned method for predicting the remaining life of a smart electric energy meter is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.
[0036] The present invention has the following beneficial effects: the present invention adaptively obtains scale parameters according to the trend of current data and the load change degree of a reference device, and predicts the remaining life of a smart electric energy meter that has not yet failed according to the adaptively obtained scale parameters, thereby improving the accuracy of the prediction result and avoiding the prediction deviation of the Weibull distribution fitting model caused by unreasonable scale parameter setting; further, the trend of current data reflects the changing trend of the value of the current data, and the larger the value, the larger the scale parameter required. The load change degree of the reference device represents the fluctuation amplitude of the current data. The fluctuation amplitude of the current data is large, that is, the larger the load change degree of the reference device, the larger the scale parameter required, and therefore obtaining the two indicators facilitates the subsequent adaptive acquisition of the scale parameter. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.
[0038] Figure 1 The present invention is a flowchart of a method for predicting the remaining life of a smart electric energy meter according to an embodiment of the present invention. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0040] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.
[0041] See also Figure 1 , which shows a flowchart of a method for predicting the remaining life of a smart electric energy meter provided by an embodiment of the present invention, the method comprising the following steps:
[0042] S001. Collect current data.
[0043] It should be noted that the current data in the smart energy meter reflects the actual working conditions and load changes of the energy meter during operation. The current data records the frequency of the meter's use, the change of current intensity, and possible load abnormalities or overloads. These factors have a great impact on the life of the energy meter. Therefore, the failed smart energy meter is used as a reference device, and the current data of the reference device is fitted with Weibull distribution to predict the remaining life of the smart energy meter that has not failed.
[0044] In the embodiment of the present invention, a failed smart energy meter of the same type and in the same use environment as the smart energy meter that has not failed is obtained from the database of the power company as a reference device, and current data within a complete use cycle of the reference device is obtained to obtain a current data sequence. It should be noted that the complete use cycle is the entire process from the first use of the reference device to its failure.
[0045] S002. Divide the current data into multiple windows, fit the current data in each window, obtain the trend characteristic value of each window according to the difference between the current data values in the window before and after fitting; obtain the similarity between adjacent windows according to the trend characteristic value of each window; obtain the trend of the current data according to the similarity between adjacent windows.
[0046] It should be noted that the Weibull distribution is a continuous probability distribution, which describes the probability density of equipment failure at a certain moment through a probability density function, and then obtains the cumulative probability of equipment failure at that moment through the cumulative distribution function of the Weibull distribution, that is, the possibility of equipment failure at that moment, to achieve the remaining life prediction of the equipment. However, in the Weibull distribution fitting model, the scale parameter determines the position and scale in the Weibull distribution, which directly affects the scaling of the probability density function of the distribution, and in practical applications, the scale parameter can reflect the life characteristics of the equipment. Therefore, by adjusting the scale parameter, the actual observed life data distribution can be better fitted, and the accuracy of the final life prediction results can be improved. In other words, adjusting the scale parameter of the Weibull distribution fitting model can improve the fitting effect of the model, thereby improving the accuracy of the remaining life prediction results of the equipment.
[0047] The distribution characteristics of known current data are usually directly related to the life of the equipment or can provide key information. At the same time, the distribution characteristics of current data can simplify the model construction and analysis process. Through appropriate analysis and fitting processes, the remaining life of the equipment can be effectively estimated without introducing complex data preprocessing or additional variables. Therefore, the scale parameters of the Weibull distribution fitting model can be adjusted based on the distribution characteristics of the current data within the complete service life of the reference equipment.
[0048] When analyzing the distribution characteristics of the current data of the reference device, it is mainly reflected from the load change degree of the reference device and the trend of the current data. This is because the trend of the current data can reflect the working status of the smart electric energy meter during actual use. With the increase of usage time, the electronic components of the electric energy meter may gradually age, resulting in fluctuations or changes in the trend of the current data. This trend change can provide an important basis for predicting the remaining life of the electric energy meter. The load change degree is an inevitable factor in the actual use of the smart electric energy meter. Due to the different electrical appliance usage of different users in different time periods, the load measured by the electric energy meter is also different. Different load conditions will have different effects on the electronic components of the electric energy meter, thereby affecting its service life. Therefore, when the load change degree of the reference device and the trend of the current data are greater, the required scale parameter is larger.
[0049] It should be further explained that the trend line obtained after fitting the current data in each window can reflect the trend of the current data. The stronger the correlation between the first-order difference of the fitted value of all current data in the window and the first-order difference of the actual value, the better the trend line reflects the trend change of the original data. In other words, its correlation can be used as a trend characteristic indicator of the data in the window, that is, the trend characteristic value of the window. When the trend characteristic values between adjacent windows are closer, it means that the adjacent windows are more similar, and the similarity between all adjacent windows is obtained. When the difference between the similarities between all adjacent windows is greater, it means that the trend of the current data is greater.
[0050] In an embodiment of the present invention, a preset window data number N is set. According to the window data number N, the current data sequence is divided into a plurality of windows in sequence. It should be noted that the number of current data in each window is N. In an embodiment of the present invention, the preset window data number N=10 is set. In other embodiments, implementers may set the value of the preset window data number N according to the specific implementation method.
[0051] Use the least square method to perform regression straight line fitting on the current data in each window, and obtain the trend line of the current data in each window and the fitting value of each current data in each window;
[0052] Obtain the absolute value of the difference between the fitted value of the nth current data in the jth window and the fitted value of the n+1th current data as the first-order difference of the fitted value of the nth current data in the jth window, and obtain the first-order difference of the fitted value of each current data in each window;
[0053] Obtain the absolute value of the difference between the actual value of the nth current data in the jth window and the actual value of the n+1th current data as the first-order difference of the actual value of the nth current data in the jth window, and obtain the first-order difference of the actual value of each current data in each window;
[0054] The DTW algorithm is used to obtain the correlation coefficient of the first-order difference of the fitted value and the first-order difference of the actual value of all current data in each window as the trend characteristic value of each window; it should be noted that the larger the value of the correlation coefficient, the larger the trend characteristic value of the window.
[0055] In the embodiment of the present invention, the similarity between the j-th window and the j+1-th window is obtained:
[0056]
[0057] In the formula, Represents the similarity between the jth window and the j+1th window; Represents the trend characteristic value of the jth window; Represents the trend characteristic value of the j+1th window; represents the slope of the trend line of the current data in the jth window; represents the intercept of the trend line of the current data in the jth window; exp() represents the exponential function with a natural constant as the base; The slope of the trend line representing the current data in the j+1th window; The intercept of the trend line representing the current data in the j+1th window; The characteristics of the trend line representing the current data in the jth window; The characteristics of the trend line representing the current data in the j+1th window; The smaller the value of is, the more similar the j-th window is to the j+1-th window.
[0058] Obtaining the trend of current data:
[0059]
[0060] In the formula, represents the trend of current data; J represents the number of windows; Represents the similarity between the jth window and the j+1th window; Represents the similarity between the j+1th window and the j+2th window; || represents the absolute value sign; the larger the value, the more obvious the trend of the current data.
[0061] S003. Obtain several maximum values in the current data; obtain the load change degree corresponding to each maximum value in the current data according to the numerical change of the maximum value; obtain the load change degree of the reference device according to the load change degree corresponding to each maximum value in the current data; obtain the scale parameter according to the load change degree of the reference device and the trend of the current data.
[0062] It should be noted that the load change degree of the reference device represents the fluctuation amplitude of the current data, which can be represented by the difference mean between all adjacent maxima of the current data. The larger its value is, the larger the fluctuation amplitude of the current data is, that is, the greater the load change degree of the reference device is. First, it is necessary to obtain the load change degree corresponding to each maximum value in the current data. Therefore, the sum of the numerical differences between any maximum value in the current data and its left and right adjacent data is taken as the numerical change of the maximum value, which describes the drastic degree of numerical change of the current data near the maximum value point. The larger the numerical change, the more drastic the corresponding current change; the sum of the numerical differences between the maximum value and the left and right adjacent maximum values is taken as the maximum value change of the maximum value, which describes the overall trend of current change between the maximum value points. The larger the maximum value change, the greater the trend; the ratio of the numerical change of the maximum value to the maximum value change reflects the significance of this change relative to the overall trend when the current changes sharply. The larger the ratio is, the more significant the load change represented by the maximum value is in the overall trend, and the greater the load change degree corresponding to the maximum value is.
[0063] In the embodiment of the present invention, a peak detection algorithm is used to obtain several maximum values in the current data.
[0064] Get the load change degree corresponding to each maximum value in the current data:
[0065]
[0066] In the formula, Represents the load change degree corresponding to the pth maximum value in the current data; represents the value of the pth maximum; Represents the value of the d+1th maximum; represents the value of the p-1th maximum; Represents the value of the current data after the pth maximum value in the current data sequence; Represents the value of the previous current data of the pth maximum value in the current data sequence; || represents the absolute value sign.
[0067] Get the load variation of the reference device:
[0068]
[0069] In the formula, Represents the degree of load change of the reference device; Represents the number of maximum values in the current data; represents the absolute value of the difference between the load change degree corresponding to the pth maximum value in the current data and the load change degree corresponding to the p+1th maximum value; The larger the value, the greater the load variation of the reference device.
[0070] It should be noted that if the current data shows an obvious trend, that is, the trend of the current data is large, a larger scale parameter is required, because the current data with a strong trend often has a larger range of variation or amplitude, and a larger scale parameter is required to effectively represent the fluctuation of the current data; if the load change of the reference device is large, it means that the fluctuation amplitude of the current data is large, that is, the difference between the maximum values of the current data is large. At this time, a larger scale parameter can usually better capture and represent this large fluctuation. Therefore, the scale parameter is obtained according to the trend of the current data and the load change of the reference device.
[0071] In an embodiment of the present invention, the scale parameter is obtained:
[0072]
[0073] In the formula, represents the scale parameter; Represents the degree of load change of the reference device; Represents the trend of current data; exp() represents an exponential function with a natural constant as the base; Represents the preset hyperparameters, which are used to The result of is constrained to [0,10] to avoid the final calculation result from taking an infinite value. In the embodiment of the present invention, the preset In other embodiments, the implementer may preset The value of Represents the base of the logarithmic function. In the embodiment of the present invention, the preset In other embodiments, the implementer may preset The value of The larger the value, the larger the scale parameter.
[0074] S004. Predict the remaining life of the smart energy meter that has not failed based on the scale parameters.
[0075] In an embodiment of the present invention, the current data of the reference device is fitted according to the acquired scale parameters to obtain a Weibull distribution model, and for the smart power meters that have not failed yet, the cumulative probability of failure of the smart power meters that have not failed yet at the current moment is obtained according to the Weibull distribution model, that is, the possibility of failure of the smart power meters that have not failed yet at the current moment, thereby realizing the prediction of the remaining life of the smart power meters that have not failed yet, helping to prevent equipment failures in advance and make effective maintenance plans.
[0076] The present invention adaptively obtains scale parameters according to the trend of current data and the load variation degree of a reference device, and predicts the remaining life of a smart electric energy meter that has not yet failed according to the adaptively obtained scale parameters, thereby improving the accuracy of the prediction result and avoiding the prediction deviation caused by unreasonable parameter setting of the Weibull distribution fitting model; further, the trend of current data reflects the changing trend of the value of the current data, and the larger the value is, the larger the scale parameter is required; the load variation degree of the reference device represents the fluctuation amplitude of the current data, and the fluctuation amplitude of the current data is relatively large, that is, the larger the load variation degree of the reference device is, the larger the scale parameter is required, and therefore obtaining the two indicator parameters facilitates the subsequent adaptive acquisition of the scale parameter.
[0077] The above-mentioned device also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, whose configuration and functions are known in the art and thus will not be described in detail here.
[0078] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high bandwidth memory, a hybrid storage cube, etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.
[0079] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.
[0080] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for predicting the remaining life of a smart electric energy meter, characterized in that: include: Collecting current data, including: obtaining current data within a complete use cycle of a failed smart energy meter of the same type and in the same use environment as a smart energy meter that has not failed, to obtain a current data sequence, wherein the complete use cycle is the entire process from the first use of the reference device to its failure; dividing the current data into multiple windows; fitting the current data in each window to obtain a fitting value of each current data in each window; obtaining a trend characteristic value of each window based on the correlation between the fitting value and the actual value of each current data in each window, including: obtaining an absolute value of a difference between a fitting value of the nth current data in the jth window and a fitting value of the n+1th current data as a first-order difference of the fitting value of the nth current data in the jth window, and obtaining a first-order difference of the fitting value of each current data in each window; Obtain the absolute value of the difference between the actual value of the nth current data in the jth window and the actual value of the n+1th current data as the first-order difference of the actual value of the nth current data in the jth window, and obtain the first-order difference of the actual value of each current data in each window; Use the DTW algorithm to obtain the first-order difference of the fitted value of all current data in each window and the correlation coefficient of the first-order difference of the actual value as the trend characteristic value of each window; obtain the similarity between adjacent windows based on the difference of the trend characteristic values of adjacent windows; obtain the trend of current data based on the similarity between adjacent windows; Obtain multiple maximum values in the current data; obtain the load change degree corresponding to each maximum value in the current data according to the numerical change of the maximum value and the current data near it; and take the average of the absolute values of the differences in the load change degrees corresponding to all adjacent maximum values as the load change degree of the reference device; Get scale parameters ; In the formula, represents the scale parameter; Represents the degree of load change of the reference device; Represents the trend of current data; exp() represents an exponential function with a natural constant as the base; represents the preset hyperparameters, Represents the base of the logarithmic function; the remaining life of the smart energy meter that has not failed is predicted based on the scale parameter.
2. The method for predicting the remaining life of a smart electric energy meter according to claim 1, characterized in that: The current data is divided into a plurality of windows, including: The number of window data N is preset, and according to the number of window data N, the current data sequence is divided into a plurality of windows in sequence, and the number of current data in each window is N.
3. The method for predicting the remaining life of a smart electric energy meter according to claim 1, characterized in that: The fitting of the current data in each window to obtain the fitting value of each current data in each window includes: The current data in each window are fitted with a regression line using the least squares method to obtain a trend line of the current data in each window and a fitting value of each current data in each window.
4. The method for predicting the remaining life of a smart electric energy meter according to claim 1, characterized in that: The obtaining of similarities between adjacent windows includes: ; In the formula, Represents the similarity between the jth window and the j+1th window; Represents the trend characteristic value of the jth window; Represents the trend characteristic value of the j+1th window; represents the slope of the trend line of the current data in the jth window; The intercept of the trend line representing the current data in the jth window; The slope of the trend line representing the current data in the j+1th window; Represents the intercept of the trend line of the current data in the j+1th window; exp() represents an exponential function with a natural constant as the base.
5. The method for predicting the remaining life of a smart electric energy meter according to claim 1, characterized in that: The obtaining of the trend of the current data includes: ; In the formula, represents the trend of current data; J represents the number of windows; Represents the similarity between the jth window and the j+1th window; Represents the similarity between the j+1th window and the j+2th window; || represents the absolute value symbol.
6. The method for predicting the remaining life of a smart electric energy meter according to claim 1, characterized in that: The obtaining of the load variation degree corresponding to each maximum value in the current data includes: ; In the formula, Represents the load change degree corresponding to the pth maximum value in the current data; represents the value of the pth maximum; represents the value of the p+1th maximum; represents the value of the p-1th maximum; Represents the value of the current data after the pth maximum value in the current data sequence; Represents the value of the previous current data of the pth maximum value in the current data sequence; || represents the absolute value sign.
7. A device for predicting the remaining life of a smart electric energy meter, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for predicting the remaining life of a smart electric energy meter according to any one of claims 1 to 6 is implemented.
