Big data-based whole life cycle management and verification method and system for electric energy meter

Abstract

A method and system for whole life cycle management and verification of an electric energy meter based on big data. A digital twin is created and maintained for each physical electric energy meter. Whole life cycle information of the electric energy meter is mapped to the digital twin. Multi-dimensional time sequence data of the digital twin is constructed and corresponding time sequence features are extracted. The time sequence features are used as a training data set to train a health degree evaluation model, and the health state of the digital twin is updated in real time according to the health degree evaluation model. Based on the updated health state of the digital twin, a dynamic health score corresponding to the digital twin is extracted, and a verification urgency index at the current time is calculated based on the change trend data of the measurement error of the electric energy meter, the severity coefficient of the operating environment of the electric energy meter and the measurement risk level of the electric energy meter. The verification urgency index is compared with the urgency threshold, and if the optimal verification time is met, a personalized verification work order is generated and pushed. The present application improves the reliability of the whole life cycle verification of the electric energy meter.

Description

Technical Field

[0001] This invention belongs to the field of electricity meter lifecycle management technology, specifically involving a method and system for electricity meter lifecycle management and verification based on big data. Background Technology

[0002] As the core equipment for measuring electricity consumption in a power system, the accuracy of electricity meters directly affects the economic operation and fair trading of the power system. To ensure the accuracy of electricity meters, relevant standards clearly require that electricity meters be regularly calibrated and that equipment with excessive metering errors be replaced or repaired in a timely manner.

[0003] Patent application CN119180211A discloses a smart meter metering and testing system and method based on digital twin technology. It collects raw data such as electrical parameters and ambient temperature from the smart meter at the site, and then performs digital modeling in the cloud to form virtual data. This allows for testing without disassembling the meter, improving the quality assessment capability throughout the smart meter's lifecycle. Patent application CN116068482A discloses a method and system for electricity metering monitoring and supplementation based on a digital twin model. It constructs a digital twin model using a mirror modeling approach, continuously inputs real-time operating data from the electricity meter, performs online identification and diagnosis of metering inaccuracies, and automatically restores the actual data after fault detection to output supplementary electricity metering data.

[0004] However, existing periodic verification models suffer from problems such as a "one-size-fits-all" approach, data gaps, delayed assessments, and crude decision-making. This invention constructs an intelligent health status assessment system for electricity meters based on full lifecycle data-driven and digital twin technology. This system enables real-time monitoring, dynamic modeling, and risk prediction of electricity meters, achieving a shift from fixed-period to dynamic intelligent verification of electricity meters throughout their entire lifecycle. Summary of the Invention

[0005] To address the problem that existing technologies cannot obtain operational data throughout the entire lifecycle of electricity meters and lack a systematic assessment of the health status of electricity meters, resulting in passive and delayed verification work, this invention provides a method and system for the full lifecycle management and verification of electricity meters based on big data.

[0006] The first aspect of this application discloses a method for the full lifecycle management and verification of electricity meters based on big data, which adopts the following technical solution: A digital twin is created and maintained for each physical electricity meter; and the full lifecycle information of the electricity meter is mapped into the digital twin; the full lifecycle information includes the static specifications and operating trajectory of the electricity meter. Construct multi-dimensional time-series data of the digital twin and extract its corresponding time-series features; use the time-series features as a training dataset to train a health assessment model, and update the health status of the digital twin in real time according to the health assessment model; Based on the updated health dynamics of the digital twin, the corresponding dynamic health score is extracted, along with the trend data of the metering error change of the electricity meter, the severity coefficient of the operating environment of the electricity meter, and the metering risk level of the electricity meter to calculate the verification urgency index at the current moment. By comparing the verification urgency index with the preset urgency threshold, if the optimal verification time is met, a personalized verification work order is generated and pushed out.

[0007] Furthermore, mapping the entire lifecycle information of the electricity meter into the digital twin includes: The static specifications include model number, rated parameters, accuracy class, manufacturing date, and hardware configuration information; At the point of time preset time period Internally, it continuously collects the operating status data and event logs of the electricity meter, and generates a dynamic time-series data stream from the operating status data; The dynamic time-series data stream is associated with the event log according to timestamps to form the running trajectory in the full life cycle information of the digital twin; The dynamic time-series data stream includes timestamps, and for each timestamp, the total positive active power, the instantaneous voltage value of each phase, the instantaneous current value of each phase, and the last power error calibration value within the preset time period corresponding to that time point.

[0008] Furthermore, the construction of the multi-dimensional time-series data of the digital twin and the extraction of its corresponding time-series features include: The multi-dimensional time-series data of the digital twin consists of dynamic time-series data streams and event logs from the electricity meter; Stability Indicators Accuracy indicators and reliability indicators The constructed comprehensive feature vector [ , , ], as the time series feature.

[0009] Furthermore, the construction of the multi-dimensional time-series data of the digital twin and the extraction of its corresponding time-series features also includes: The stability index is obtained by weighting the voltage variation coefficient and current fluctuation rate within a preset time period corresponding to each time point in the dynamic time-series data stream. The voltage variation coefficient is calculated as follows: within the dynamic time-series data stream, based on the standard deviation of each element in the set of instantaneous voltage values ​​within a preset time period corresponding to each time point. with the mean The ratio of the two values ​​is used to obtain the voltage variation coefficient; The current fluctuation rate is calculated as follows: within the dynamic time-series data stream, based on the set of instantaneous current values ​​within the preset time period corresponding to each time point, the maximum instantaneous current value in the set is taken. Minimum instantaneous current value The difference between the two and the average instantaneous current value in the set. The ratio of the two values ​​is used to obtain the current fluctuation rate. The voltage variation coefficient weight Weighted by current volatility The sum is 1, and >0, >0.

[0010] Further, the accuracy indicators are extracted, including: The difference between the total positive active power at each point in time within the preset time period and the corresponding actual energy value measured by the standard meter or the actual power generation is the error calibration value. The error calibration values ​​at each time point within the preset calibration period are used to construct an error calibration value sequence in chronological order. Calculate the mean of the absolute values ​​of all elements in the sequence. The sequence was then linearly fitted to obtain its linear regression slope. and the calculated and We then apply weights to obtain an accuracy index.

[0011] Further, the reliability indicators are extracted, including: The reliability index is obtained by weighting the communication success rate within a preset time period corresponding to each time point in the dynamic time-series data stream with the comprehensive risk of various events in the event log. The calculation method for the comprehensive risk is as follows: based on the preset time period corresponding to each time point, the event log of the electricity meter will be used to calculate the risk level. This type of event is based on the frequency of its occurrence within a preset time period from that point in time. The severity level is determined by the score corresponding to the degree of impact of this type of event. Multiply the two to obtain the risk of a single event category; sum the risks of all event categories to obtain the comprehensive risk.

[0012] Furthermore, the step of training the health assessment model using time-series features as the training dataset includes: The time series features are preprocessed by normalization and divided into training and validation sets; Initialize the parameters of the random forest model, iteratively train the health assessment model using the training set, and adjust the hyperparameters of the health assessment model using the validation set; Training ends and the health status of the health assessment model is output when the prediction error of the health assessment model is lower than a preset threshold.

[0013] Furthermore, the calculation of the urgency index of the current test includes: The urgency index of the current energy meter's verification The calculation method is as follows: the dynamic health score, error change trend data, severity coefficient of the operating environment and measurement risk level are weighted and summed to obtain the verification urgency index; The dynamic health score is obtained based on the health status of the electricity meter through a corresponding relationship; The error change trend data is obtained by performing linear regression analysis on the error calibration value sequence of the electricity meter in the historical time period, calculating the slope of the regression line, and using the absolute value of the slope as the error change trend data. The environmental severity coefficient is obtained by querying the environmental status form of the electricity meter. The metering risk level is obtained by querying the electricity meter in a pre-set form in the database.

[0014] Furthermore, the step of comparing the verification urgency index with a preset urgency threshold, and generating a personalized verification work order if the optimal verification time is met, includes: The urgency index of the test Compared with the preset first urgency threshold Compare; like If the current time is within the optimal verification time, the meter is retained; otherwise, it is discarded. Generate personalized verification work orders for electricity meters that have reached the optimal verification time.

[0015] Furthermore, the generation of personalized verification work orders includes: The personalized verification work order includes the electricity meter's serial number, installation address, verification urgency index, and recommended verification date; The recommended verification date is equal to the current time. The fault tolerance interval for the electricity meter sum; Calculate the fault tolerance interval for the electricity meter. The process is as follows: Extract the urgency index of the last verification before the meter's verification result becomes abnormal. And from the urgency index of this test The duration between the time of the test result and the time of the abnormal result is used as the similarity test interval. ; Calculate the urgency index of the electricity meter's verification at the current moment. and absolute difference ; Will Compared with the preset second urgency threshold Comparison, for Historical data will be retained, and all historical data that meet the conditions will be included. Form a set S; Calculate all of set S The average value is the calibration tolerance interval of the electricity meter. ; The abnormal test result refers to a health score output by the health assessment model that is lower than the preset value, or an abnormality in the actual health status of the electricity meter as determined by laboratory testing or on-site testing.

[0016] The second aspect of this application discloses a big data-based system for the full lifecycle management and verification of electricity meters, implementing the technical solution of the electricity meter full lifecycle management and verification method described in the first aspect of this application. The system includes: The digital twin management module creates and maintains a digital twin for each physical electricity meter; and maps the electricity meter's full lifecycle information into the digital twin; the full lifecycle information includes the electricity meter's static specifications and operating trajectory; The health assessment module constructs multi-dimensional time-series data of the digital twin and extracts its corresponding time-series features; it uses the time-series features as a training dataset to train the health assessment model and updates the health status of the digital twin in real time based on the health assessment model. The verification and push module, based on the updated health dynamics of the digital twin, extracts the corresponding dynamic health score, the trend data of the metering error change of the electricity meter, the severity coefficient of the operating environment of the electricity meter, and the metering risk level of the electricity meter to calculate the verification urgency index at the current moment; by comparing the verification urgency index with the preset urgency threshold, if the optimal verification time is met, a personalized verification work order is generated and pushed.

[0017] The beneficial effects of this invention are that, compared with the prior art, (1) This invention achieves a fundamental change in the management mode of electricity meters by creating a digital twin of the electricity meter, constructing a health assessment model, and using a dynamic decision-making algorithm based on multi-factor fusion; to a certain extent, it abandons the rigid fixed cycle mode and makes precise intervention based on the real-time individual health status of each electricity meter, which greatly improves the pertinence of operation and maintenance work.

[0018] (2) This invention constructs and continuously maintains a full life cycle digital twin for each electricity meter, which carries static specifications, dynamic time-series data streams and health status. Through a real-time data interaction interface, it realizes the synchronous mapping between physical entities and virtual models, providing a data foundation for subsequent refined evaluation and dynamic decision-making. (3) The present invention constructs a health assessment model based on random forest. By extracting stability indicators, accuracy indicators and reliability indicators from multi-dimensional time series data to form a comprehensive feature vector, the abstract health status of the electricity meter is quantified into a specific dynamic health score, making the health status traceable and providing accurate data support for subsequent verification decisions. (4) The present invention constructs a verification priority assessment model, which integrates multiple factors such as dynamic health score, measurement error change trend, environmental severity parameter and measurement risk level, dynamically calculates the verification urgency index, and combines the verification fault tolerance interval time in historical data to accurately generate personalized verification work orders. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the big data-based method for the full lifecycle management and verification of electricity meters according to the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this invention are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0021] Existing technologies have many inherent defects and are no longer able to meet the refined requirements of electricity meter metering management for the large-scale development of smart grids. Current periodic verification methods generally employ fixed verification cycles, failing to consider the differences in the operating status of individual electricity meters. For example, some electricity meters in good operating condition and with stable errors are prematurely replaced or removed for verification, resulting in a significant waste of human and material resources. Conversely, some electricity meters operating in harsh environments or with early-stage potential faults may not receive timely intervention due to not having reached their verification cycle, potentially leading to metering disputes or safety accidents. Furthermore, current technology only performs a one-time test on basic parameters such as voltage, current, and power at the verification point, failing to obtain operational data throughout the meter's entire lifecycle. It lacks a systematic assessment method for the health status of electricity meters and cannot predict potential fault risks in advance, leading to passive and delayed verification work. Moreover, current verification work relies heavily on manual recording and analysis, failing to establish an effective connection between physical electricity meters and virtual models, thus hindering real-time mapping and dynamic simulation of operational data. The determination of verification time is mainly based on experience or standard regulations, without comprehensive judgment considering the actual health status of the electricity meters, environmental severity, and metering risk level, resulting in a lack of scientific basis for verification decisions and difficulty in achieving optimal resource allocation.

[0022] This invention aims to address technical biases such as resource misallocation and decision-making lag in existing technologies, and provides a fundamental solution from "periodic general inspection" to "on-demand precise inspection", which significantly improves the targeted nature of operation and maintenance, saves manpower and material resources, and enhances the reliability of metering.

[0023] This embodiment uses a municipal power supply company as an example to implement the method of the present invention. A central server (for digital twin management, model calculation and decision-making), an edge computing gateway (for distribution area data aggregation and lightweight computing), and a mobile application terminal (for work order reception and execution) are deployed on a cloud platform.

[0024] As an embodiment of the present invention, a specific implementation method for the full lifecycle management and verification of electricity meters based on big data is disclosed. The execution flow of the method embodiment is as follows: Figure 1 .

[0025] S1: Create and maintain a corresponding digital twin for each electricity meter. Configure a real-time data interaction interface for the digital twin. Through the real-time data interaction interface, complete the real-time data mapping of the electricity meter in the corresponding digital twin and reproduce the full life cycle information of the corresponding electricity meter in the virtual space. The full life cycle information includes the static specifications, operating trajectory and health status of each electricity meter.

[0026] S1.1: Collect the static specifications of each electricity meter, such as meter model, rated voltage, rated current, accuracy class, manufacturing date and hardware configuration information.

[0027] S1.2: Continuously acquire the operating status data and event logs of the physical electricity meter through the communication network, and synchronously update them to the operating trajectory of its digital twin.

[0028] Specifically, based on the electricity meter at a certain point in time preset time period (like The operational status data collected within the last 15 minutes is used as the dynamic time-series data stream of this digital twin; The dynamic time-series data stream includes timestamps, and for each timestamp, the total positive active power, the instantaneous voltage value of each phase, the instantaneous current value of each phase, and the last error calibration value; the event log includes clock battery undervoltage, overvoltage, current loss, reverse phase sequence events, and occurrence events; the last error calibration value is dynamically updated and is obtained by the energy meter itself completing error assessment and calibration calculation locally and actively reporting the latest calibration results.

[0029] The dynamic time-series data stream of the digital twin corresponding to a certain timestamp and the event log corresponding to that timestamp together constitute multi-dimensional time-series data.

[0030] The running trajectory consists of a dynamic time-series data stream and an event log, and is a full lifecycle state evolution process that is associated and sorted according to timestamps.

[0031] S1.3: Associate the dynamic time-series data stream of the digital twin corresponding to each timestamp with the event log of the electricity meter according to the timestamp, and update the running trajectory of the corresponding digital twin.

[0032] In this embodiment, a hybrid storage approach combining a time-series database (InfluxDB) and a relational database (MySQL) is used. Static specification parameters are stored in MySQL, while dynamic time-series data streams are stored in InfluxDB.

[0033] S2: For each electricity meter, by constructing a real-time data interaction interface for the corresponding digital twin, multi-dimensional time-series data during the operation of the electricity meter is collected and input into a pre-trained health assessment model. The health assessment model outputs the dynamic health score of the digital twin and updates the health status of the digital twin in real time.

[0034] In this embodiment, the health assessment model adopts a random forest model. During the health status assessment process based on the health assessment model, time series features of multi-dimensional time series data are extracted, and health scores are calculated based on the extracted time series features and the static specification parameters of the digital twin. During the model training process, group modeling training is performed through static specification parameters, with each group corresponding to a model to be trained. However, the input of the grouped models during the training process is the time series features; that is, the matching models are first selected based on the static specification parameters, and then the time series features are input into the selected matching models to obtain the corresponding health scores.

[0035] S2.1: Extract time series features from multi-dimensional time series data. The time series features are a comprehensive feature vector composed of stability indicators, accuracy indicators, and reliability indicators.

[0036] S2.1.1: Stability index, obtained by calculating the voltage variation coefficient and current fluctuation rate in the dynamic time series data stream of multi-dimensional time series data at a certain point in time.

[0037] Furthermore, stability indicators in time series features Calculation formula: ; in, Represents the volatility function. This represents the coefficient of variation function, which is equal to the ratio of the standard deviation to the mean of each element in the set corresponding to the independent variable. For example, let the independent variable be... The corresponding set ; and These represent the weights for voltage variation coefficient and current fluctuation rate, respectively. + =1; In this embodiment =0.5, =0.5.

[0038] ; ; ; in, Standard deviation; The mean; It is a set The mean, For set The first in There are 10 data points; n is the number of elements in the set. For the data points in the set, =1,2,3,…,n.

[0039] The It represents the set of instantaneous voltage values ​​within a dynamic time-series data stream; It represents the set of instantaneous current values ​​within a dynamic time-series data stream; The current fluctuation rate is represented and its calculation method is described below: Let the sequence of instantaneous current values ​​corresponding to IR be { }, =1,2,3,…,n, To label the data points in the sequence, The first in the sequence of instantaneous current values One data point; but Among them, the maximum instantaneous current value =max({ Minimum instantaneous current value =min({ }); Average instantaneous current value .

[0040] S2.1.2: Accuracy index, calculated based on the error calibration value sequence within the dynamic time series data stream in multi-dimensional time series data at a certain point in time.

[0041] The error calibration value refers to the relative deviation between the energy value recorded by the energy meter (the total positive active energy of the dynamic time-series data stream) and the standard energy value (the energy value actually measured by the standard meter or the actual power generation); the calculation method is: (energy meter reading - standard energy value) / standard energy value × 100%.

[0042] First, this embodiment assumes that the error was calibrated once a month in the past. Therefore, the error calibration value sequence formed by the past twelve calibration points is as follows: [+0.10%,+0.11%,+0.12%,+0.14%,+0.16%,+0.18%,+0.21%,+0.23%,+0.26%,+0.28%,+0.31%,+0.35%]; the +0.10% in the sequence indicates that the electricity meter recorded 0.10% more electricity than the actual amount consumed. Furthermore, calculate the mean of the absolute values ​​of each element in the sequence. and the linear regression slope of the sequence Accuracy indicators The calculation formula is: ; in, and These represent the preset third and fourth weight values, respectively. + =1.

[0043] S2.1.3: Reliability index, calculated based on the communication success rate and frequency of various events within the dynamic time-series data stream in multi-dimensional time-series data at a certain point in time.

[0044] Reliability indicators The calculation formula is: ; in, Indicates the communication success rate; This indicates the number of event types in the event log. Indicates the type of event in the event log. =1,2,3,…,N; Indicates the first The frequency of occurrence of this type of event within a preset period; Indicates the first Preset severity levels for this type of event; and These represent the preset fifth and sixth weight values, respectively. + =1; In this embodiment =0.6, =0.4.

[0045] Preset severity level The scores are quantitative values ​​set according to the degree of impact of the event. The higher the value, the higher the level. Each level is assigned a specific weight or value, such as Level 1 (Severe) is 1.0, Level 2 (Moderate) is 0.5, and Level 3 (Minor) is 0.1.

[0046] S2.2: Pre-training process of health assessment model.

[0047] S2.2.1: Collect the full life cycle operation data and corresponding actual health status assessment results of multiple physical energy meters of the same type to construct a training dataset.

[0048] The full life cycle operation data refers to the multi-dimensional time-series data generated during the entire process of the electricity meter from installation and commissioning to scrapping and decommissioning; the actual health status assessment result refers to an authoritative qualitative or quantitative evaluation of the true health status of the electricity meter at a certain point in time. The actual health status assessment results are based on expert experience or historical operation and maintenance records, classifying electricity meters into levels such as excellent, good, average / sub-healthy, and faulty.

[0049] S2.2.2: Normalize the multi-dimensional time-series data in the obtained training dataset and divide it into training and validation sets.

[0050] Stability Indicators Accuracy indicators and reliability indicators The resulting comprehensive feature vector serves as the feature of the time series. This comprehensive feature vector undergoes normalization preprocessing to eliminate differences between different units.

[0051] The normalized comprehensive feature vector is used as the input feature of the random forest model, and together with its corresponding actual health status assessment results, forms the training sample set. The training sample set is then divided into a training set and a validation set according to a preset ratio.

[0052] S2.2.3: Initialize the parameters of the random forest model, train the health assessment model using the training set, and adjust the hyperparameters of the health assessment model using the validation set until the prediction error of the health assessment model is lower than the preset threshold, thus completing the pre-training of the health assessment model.

[0053] The training set is used to iteratively update the parameters of the random forest model, and the validation set is used to monitor the model training effect and adjust the hyperparameters until the model prediction error is lower than the preset threshold, thus completing the model training.

[0054] The pre-trained health assessment model processes the extracted time-series features of the multi-dimensional time-series data and outputs the original assessment score of the multi-dimensional time-series data. Then, using a preset normalization function, it maps the original assessment score of the multi-dimensional time-series data to the [1, 100] interval to obtain the dynamic health score corresponding to the multi-dimensional time-series data. ; The calculation formula is as follows: ; Wherein, SY represents the original evaluation score of the multi-dimensional time series data; express Mapped values ​​in the range [1, 100]; This represents the minimum original assessment score predicted by the health assessment model on the training set; This represents the maximum raw assessment score predicted by the health assessment model on the training set.

[0055] Health status and health score are corresponding. In this embodiment, health status includes excellent, good, average / sub-healthy and faulty. The corresponding relationship between health status and health score is as follows: excellent is 90-100 points, good is 75-89 points, average / sub-healthy is 60-74 points and faulty is 0-59 points.

[0056] S3: Retrieve the updated digital twin information from S2, extract the corresponding dynamic health score and measurement error change trend data; obtain the severity parameters of the operating environment of the electricity meter through the environmental data acquisition interface of the digital twin; combine the preset electricity meter measurement risk level classification standard, construct a verification priority evaluation model based on dynamic health score, measurement error change trend, and environmental severity parameters, dynamically calculate the optimal verification time for the electricity meter corresponding to the digital twin, and generate and push personalized verification work orders.

[0057] S3.1: Dynamically calculate the urgency index of the electricity meter's verification at the current moment.

[0058] The urgency index is determined using the following formula: ; in, For the first The urgency index of the current energy meter's verification; Indicates the first The dynamic health score of each electricity meter at the current moment; Indicates the first The absolute value of the error change trend of each electricity meter at the current moment, i.e., the slope of the linear regression. To quantify; Indicates the first The environmental severity coefficient of each electricity meter at the current moment. The value is obtained by querying the environmental severity coefficient corresponding to the environmental monitoring data in the preset energy meter environmental status form. This form is a manually set form, and the environmental monitoring data includes temperature, humidity and salt spray concentration. Indicates the first The metering risk level coefficient of each electricity meter in the database preset form; , , and These represent the weighting coefficients for dynamic health score, error change trend, environmental severity, and measurement risk level, respectively. .

[0059] The environmental severity levels are divided into: mild environment, general environment, and severe environment; the corresponding humidity for a mild environment is... 75%, the typical humidity level in a given environment. 95%, humidity level suitable for harsh environments 95%.

[0060] The metering risk level coefficient refers to a pre-set quantitative weight value based on the technical classification of the electricity metering device, the importance of the application scenario, and the user type. The corresponding risk levels are divided into: extremely high risk, high risk, medium risk, low risk, and extremely low risk. The metering risk level corresponds to the device category: extremely high risk corresponds to category I, high risk corresponds to category II, medium risk corresponds to category III, low risk corresponds to category IV, and extremely low risk corresponds to category V.

[0061] S3.2: Verification urgency index calculated based on electricity meter readings Compared with the preset first urgency threshold Compare them.

[0062] like If the current time is not reached, the energy meter is determined to have not reached the optimal calibration time; otherwise, the energy meter is determined to have reached the optimal calibration time.

[0063] S3.3: Summarize the energy meters that have reached the optimal verification time at the current time to obtain the personalized verification work order set corresponding to the current time.

[0064] Furthermore, each element in the set of personalized verification work orders corresponding to the current time corresponds to a personalized verification work order for an electricity meter; the personalized verification work order includes the corresponding electricity meter number, installation address, verification urgency index, and recommended verification date; the recommended verification date is equal to the sum of the current time and the verification fault tolerance interval of the corresponding electricity meter.

[0065] In a further implementation, the recommended calibration date of the electricity meter is obtained.

[0066] S3.3.1: Extract the verification urgency index of the electricity meter at the corresponding installation location from the historical data before the verification result is abnormal. And the length of similar testing intervals from the time the index is calculated to the time when the test result is abnormal. .

[0067] Abnormal test results refer to situations where the pre-trained health assessment model outputs a health score lower than a preset value, or where the actual health status of the electricity meter is abnormal as determined by laboratory testing or on-site testing.

[0068] S3.3.2: Calculate the urgency index of the current electricity meter at the current moment. Urgency index of the last test before the test result was abnormal absolute difference .

[0069] S3.3.3: For Historical data is retained, and the corresponding similarity verification intervals in all historical data are recorded. Form a set S; where, This refers to the preset second urgency threshold.

[0070] S3.3.4: If set S is not empty, then calculate the average similarity test interval duration in set S. This refers to the calibration tolerance interval of the corresponding electricity meter.

[0071] Verification tolerance interval duration The calculation formula is: ; in, Let S be the number of elements in set S; if set S is empty, This is the default value that is set.

[0072] The verification tolerance interval time obtained through the above steps Combined with the current time Recommended testing date can be obtained. .

[0073] S3.4: Directly push the installation addresses of each electricity meter involved in the personalized verification work order set corresponding to the current time to the verification personnel in the respective management area, track the execution status of the work orders, and feed the results back to the corresponding digital twin.

[0074] As an embodiment of this application, a big data-based full lifecycle management and verification system for electricity meters is disclosed. Employing the specific implementation method described above for the full lifecycle management and verification of electricity meters, the system includes: The digital twin management module creates and maintains a digital twin for each physical electricity meter; and maps the electricity meter's full lifecycle information into the digital twin; the full lifecycle information includes the electricity meter's static specifications and operating trajectory; The health assessment module constructs multi-dimensional time-series data of the digital twin and extracts its corresponding time-series features; it uses the time-series features as a training dataset to train the health assessment model and updates the health status of the digital twin in real time based on the health assessment model. The verification and push module, based on the updated health dynamics of the digital twin, extracts the corresponding dynamic health score, the trend data of the metering error change of the electricity meter, the severity coefficient of the operating environment of the electricity meter, and the metering risk level of the electricity meter to calculate the verification urgency index at the current moment; by comparing the verification urgency index with the preset urgency threshold, if the optimal verification time is met, a personalized verification work order is generated and pushed.

[0075] As an embodiment of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, it adopts the specific implementation method described above for the full life cycle management and verification method of the electricity meter.

[0076] As an embodiment of the present invention, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, it adopts the specific implementation method described above for the full life cycle management and verification method of the electricity meter.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for full lifecycle management and verification of electricity meters based on big data, characterized in that, The method includes: A digital twin is created and maintained for each physical electricity meter; and the full lifecycle information of the electricity meter is mapped into the digital twin; the full lifecycle information includes the static specifications and operating trajectory of the electricity meter. Construct multi-dimensional time-series data of the digital twin and extract its corresponding time-series features; use the time-series features as a training dataset to train a health assessment model, and update the health status of the digital twin in real time according to the health assessment model; The multi-dimensional time-series data of the digital twin consists of dynamic time-series data streams and event logs from the electricity meter; Stability Indicators Accuracy indicators and reliability indicators The constructed comprehensive feature vector [ , , ], as the time series feature; The stability index is obtained by weighting the voltage variation coefficient and current fluctuation rate within a preset time period corresponding to each time point in the dynamic time-series data stream. The difference between the total positive active power at each point in time within the preset time period and the corresponding actual energy value measured by the standard meter or the actual power generation is the error calibration value. The error calibration values ​​at each time point within the preset calibration period are used to construct an error calibration value sequence in chronological order. Calculate the mean of the absolute values ​​of all elements in the sequence. And perform linear fitting on the sequence to obtain its linear regression slope. and the calculated and We then apply weighted averages to obtain an accuracy index. The reliability index is obtained by weighting the communication success rate within a preset time period corresponding to each time point in the dynamic time-series data stream with the comprehensive risk of various events in the event log. Based on the updated health dynamics of the digital twin, the corresponding dynamic health score, the trend data of the metering error, the severity coefficient of the operating environment of the meter, and the metering risk level are extracted to calculate the verification urgency index at the current moment. The verification urgency index is calculated using the following formula: ； in, For the first The urgency index of the current energy meter's verification; Indicates the first The dynamic health score of each electricity meter at the current moment; Indicates the first The absolute value of the error change trend of each electricity meter at the current moment, i.e., the slope of the linear regression. To quantify; Indicates the first The environmental severity coefficient of each electricity meter at the current moment. The value is obtained by querying the environmental severity coefficient corresponding to the environmental monitoring data in the preset electricity meter environmental status form; Indicates the first The metering risk level coefficient of each electricity meter in the database preset form; , , and These represent the corresponding weights, and the sum of the weights is 1. By comparing the verification urgency index with a preset urgency threshold, if the optimal verification time is met, a personalized verification work order is generated and pushed out; the recommended verification date in the personalized verification work order is obtained as follows: Extract the urgency index of the last inspection of the electricity meter before the inspection result is abnormal. And from the urgency index of this test The duration between the time of the test result and the time of the abnormal result is used as the similarity test interval. ; Calculate the urgency index of the electricity meter's verification at the current moment. and absolute difference ;Will Compared with the preset second urgency threshold Comparison, for The tolerance interval for testing is obtained by averaging the durations of all similar testing intervals in the historical data. ; set the current time The fault tolerance interval for the electricity meter The sum is used as the recommended verification date.

2. The method for full lifecycle management and verification of electricity meters based on big data as described in claim 1, characterized in that, The process of mapping the entire lifecycle information of the electricity meter into the digital twin includes: The static specifications include model number, rated parameters, accuracy class, manufacturing date, and hardware configuration information; At the point of time preset time period Internally, it continuously collects the operating status data and event logs of the electricity meter, and generates a dynamic time-series data stream from the operating status data; The dynamic time-series data stream is associated with the event log according to timestamps to form the running trajectory in the full life cycle information of the digital twin; The dynamic time-series data stream includes timestamps, and for each timestamp, the total positive active power, the instantaneous voltage value of each phase, the instantaneous current value of each phase, and the last power error calibration value within the preset time period corresponding to that time point.

3. The method for full lifecycle management and verification of electricity meters based on big data as described in claim 1, characterized in that, The construction of the digital twin's multi-dimensional time-series data and extraction of its corresponding time-series features includes: The voltage variation coefficient is calculated as follows: within the dynamic time-series data stream, based on the standard deviation of each element in the set of instantaneous voltage values ​​within the preset time period corresponding to each time point. with the mean The ratio of the two values ​​is used to obtain the voltage variation coefficient; The current fluctuation rate is calculated as follows: within the dynamic time-series data stream, based on the set of instantaneous current values ​​within the preset time period corresponding to each time point, the maximum instantaneous current value in the set is taken. Minimum instantaneous current value The difference between the two and the average instantaneous current value in the set. The ratio of the two values ​​is used to obtain the current fluctuation rate. The voltage variation coefficient weight Weighted by current volatility The sum is 1, and >0, >

0.

4. The method for full lifecycle management and verification of electricity meters based on big data as described in claim 1, characterized in that, Extracting the reliability metrics includes: The comprehensive risk is calculated as follows: based on the preset time period corresponding to each time point, the event log of the electricity meter is used to calculate the risk level. This type of event is based on the frequency of its occurrence within a preset time period from that point in time. The severity level is determined by the score corresponding to the degree of impact of this type of event. Multiply the two to obtain the risk of a single event category; sum the risks of all event categories to obtain the comprehensive risk.

5. The method for full lifecycle management and verification of electricity meters based on big data as described in claim 1, characterized in that, The step of training a health assessment model using time-series features as a training dataset includes: The time series features are preprocessed by normalization and divided into training and validation sets; Initialize the parameters of the random forest model, iteratively train the health assessment model using the training set, and adjust the hyperparameters of the health assessment model using the validation set; Training ends and the health status of the health assessment model is output when the prediction error of the health assessment model is lower than a preset threshold.

6. The method for full lifecycle management and verification of electricity meters based on big data as described in claim 1, characterized in that, The calculation of the urgency index of the current verification time includes: The urgency index of the current energy meter's verification The calculation method is as follows: the dynamic health score, error change trend data, severity coefficient of the operating environment and measurement risk level are weighted and summed to obtain the verification urgency index; The dynamic health score is obtained based on the health status of the electricity meter through a corresponding relationship; The error change trend data is obtained by performing linear regression analysis on the error calibration value sequence of the electricity meter in the historical time period, calculating the slope of the regression line, and using the absolute value of the slope as the error change trend data. The environmental severity coefficient is obtained by querying the environmental status form of the electricity meter. The metering risk level is obtained by querying the electricity meter in a pre-set form in the database.

7. The method for full lifecycle management and verification of electricity meters based on big data according to claim 1, characterized in that, The process involves comparing the verification urgency index with a preset urgency threshold. If the optimal verification time is met, a personalized verification work order is generated, including: The urgency index of the test Compared with the preset first urgency threshold Compare; like If the current time is within the optimal verification time, the meter is retained; otherwise, it is discarded. Generate personalized verification work orders for electricity meters that have reached the optimal verification time.

8. The method for full lifecycle management and verification of electricity meters based on big data according to claim 7, characterized in that, The generation of personalized verification work orders includes: The personalized verification work order also includes the meter number, installation address, and verification urgency index.

9. A big data-based electricity meter lifecycle management and verification system, executing the electricity meter lifecycle management and verification method as described in any one of claims 1-8, characterized in that, The system includes: The digital twin management module creates and maintains a digital twin for each physical electricity meter; and maps the electricity meter's full lifecycle information into the digital twin; the full lifecycle information includes the electricity meter's static specifications and operating trajectory; The health assessment module constructs multi-dimensional time-series data of the digital twin and extracts its corresponding time-series features; it uses the time-series features as a training dataset to train the health assessment model and updates the health status of the digital twin in real time based on the health assessment model. The verification and push module, based on the updated health dynamics of the digital twin, extracts the corresponding dynamic health score, the trend data of the metering error change of the electricity meter, the severity coefficient of the operating environment of the electricity meter, and the metering risk level of the electricity meter to calculate the verification urgency index at the current moment; by comparing the verification urgency index with the preset urgency threshold, if the optimal verification time is met, a personalized verification work order is generated and pushed.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is loaded into the processor, it implements the method for full life cycle management and verification of electricity meters according to any one of claims 1-8.

11. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for full life cycle management and verification of electricity meters according to any one of claims 1-8.

Images