Method and system for dynamic adjustment of calibration period for electric energy meter

Abstract

This invention discloses a method and system for dynamically adjusting the calibration cycle of electricity meters, relating to the field of smart grid electricity metering technology. The invention first uses an edge computing module to receive multiple self-diagnostic signals uploaded by the electricity meter. After judging the validity of the signals, it calculates the comprehensive degradation trend index and its rate of change. Based on the rate of change, it generates calibration cycle compression or extension instructions in real time. Simultaneously, an incremental proportional-integral-derivative controller corrects the adjustment strategy based on measured error feedback. The intermediate correction data is then uploaded to the cloud-based metering data center, where the cloud updates the degradation model parameters for the same batch and issues the long-term baseline calibration cycle. This invention enables closed-loop dynamic adaptive adjustment of the electricity meter calibration cycle, effectively improving the reliability of electricity metering, reducing metering operation and maintenance costs, and is applicable to the full lifecycle operation and maintenance management of smart grid electricity meters.

Description

Technical Field

[0001] This invention relates to the field of smart grid power metering technology, specifically to a method and system for dynamically adjusting the calibration cycle of power meters. Background Technology

[0002] With the continuous development of smart grid construction, electricity meters, as core metering instruments for trade settlement, directly impact the economic interests of both power supply companies and electricity customers. According to current metrological regulations, electricity metering instruments need to be calibrated periodically to ensure the traceability of their values. However, with the continuous growth in the number of installed smart meters—more than 600 million have been deployed nationwide to date—the traditional fixed-period rotation or calibration model is facing multiple challenges, including significant resource waste, high maintenance costs, and the impact of power outages on users. How to achieve a scientific, differentiated, and dynamic adjustment of the electricity meter calibration cycle while ensuring metering accuracy has become a pressing technical problem for the industry.

[0003] Several existing technologies provide solutions for adjusting the calibration cycle or online calibration of electricity meters. For example, invention patent publication number CN118534405A discloses a calibration method for electricity meters. This method obtains the historical installation and maintenance records of all electricity meters and calculates the deviation correlation coefficient; it obtains the daily deviation of each maintenance cycle and calculates the real-time calibration cycle of the electricity meter; it obtains the usage time and total measured electricity of new meters without maintenance records, and calculates the real-time error rate of the new meter by combining the deviation correlation coefficient. The core of this solution is to use external operating data such as historical maintenance records and daily deviation to calculate the calibration cycle. Essentially, it is an open-loop adjustment method based on historical statistical data. Once the calibration cycle is calculated and determined, it will not change before the next calculation cycle, and it lacks real-time response capability to sudden error drift. For example, the invention patent publication number CN118688709A discloses an adaptive calibration method and system for electricity meters. This method includes acquiring basic and supplementary data from the electricity meter, performing time-series data alignment, periodically dividing the environmental dataset, generating time-series segmentation results, acquiring the meter's device information, performing key feature analysis within the period, establishing interference weight factors for each feature, and constructing a performance degradation model. Although this scheme introduces environmental data and time-series analysis, its decision-making logic remains a one-way open-loop model of "periodic evaluation—outputting periodic recommendations." There is a lack of continuous, dynamic closed-loop interaction between the data used for state evaluation and the calibration cycle adjustment decisions, and it does not involve real-time perception of deeper degradation signals within the metering chip.

[0004] Furthermore, regarding self-testing technology for metering chips, invention patent publication number CN113917385B discloses a self-testing method and system for electricity meters. The MCU used in this system features flash memory ECC error correction and detection, and the RAM has real-time byte parity checking, internal oscillator clock stop detection, and write protection for key registers. This solution utilizes some of the chip's internal detection functions to achieve fault self-testing, but its purpose is to determine whether the chip has experienced a functional failure, rather than assessing the degree of progressive performance degradation, and it does not establish a correlation between the chip's internal self-diagnostic signals and calibration cycle adjustment decisions. In terms of chip reference voltage monitoring, some patents disclose schemes that can accurately determine whether the first reference source is faulty or whether the first reference voltage it generates has experienced voltage drift, but their purpose is also limited to chip-level fault diagnosis and is not used for triggering and adjusting calibration cycles.

[0005] In summary, existing technologies share a common and fundamental technical deficiency: the calibration cycle adjustment system lacks a complete technical solution that can simultaneously acquire early degradation signals from within the metering chip, dynamically adjust the calibration cycle in real time based on these signals, and continuously optimize the adjustment strategy through closed-loop feedback. Specifically, existing solutions either rely on external operational data (such as daily deviations, historical maintenance records, and environmental data) for coarse-grained cycle calculations, lacking the ability to detect early abnormal signals occurring within the metering chip, such as reference source drift and ADC gain degradation; or, while utilizing some of the metering chip's self-detection functions, they only use them for fault alarms rather than calibration cycle decisions. More importantly, existing solutions generally lack the technical means to implement the complete closed-loop chain of "monitoring—adjustment—verification—correction strategy—re-monitoring," resulting in the calibration cycle adjustment strategy failing to self-optimize based on actual performance. Therefore, the field needs a technical solution that organically integrates the perception of self-diagnostic signals within the metering chip, multi-timescale cycle-level decision-making, and closed-loop feedback strategy self-optimization to achieve a transformation of the electricity meter calibration cycle from "static allocation" to "dynamic adaptation." Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for dynamically adjusting the calibration cycle of electricity meters. By reading multiple self-diagnostic signals from the metering chip of the electricity meter, a closed-loop dynamic adjustment mechanism on the edge side and an end-edge-cloud collaborative architecture are constructed. This can accurately capture early performance degradation signs of the metering chip, realize dynamic adaptive optimization of the electricity meter calibration cycle, effectively ensure metering accuracy, and reduce the overall cost of power grid metering operation and maintenance.

[0007] In a first aspect, to solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for dynamically adjusting the calibration cycle of an electricity meter, the method being executed by an edge computing module deployed in a distribution concentrator or edge computing gateway, comprising:

[0008] The device receives six types of self-diagnostic signals uploaded by the energy meter. These six types of self-diagnostic signals are output in real time by the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register, and clock frequency deviation register inside the energy meter's metering chip.

[0009] After judging the validity of the six types of self-diagnostic signals received, the comprehensive degradation trend index is calculated, and the rate of change of the comprehensive degradation trend index within the preset sliding time window is calculated.

[0010] When the rate of change of the comprehensive degradation trend index exceeds the preset warning threshold, the calibration cycle compression command is automatically generated without waiting for the current calibration cycle to end. The remaining calibration cycle of the energy meter is compressed to a set proportion of the original remaining calibration cycle according to the preset mapping relationship. When the absolute value of the rate of change of the comprehensive degradation trend index is lower than the preset stability threshold in multiple consecutive monitoring windows, the calibration cycle extension command is automatically generated to extend the remaining calibration cycle of the energy meter according to the preset extension coefficient.

[0011] The compression range of the calibration cycle compression command and the extension range of the calibration cycle extension command are compared with the measured error value obtained in the subsequent actual calibration of the energy meter to calculate the prediction deviation. The incremental proportional-integral-derivative controller is used to correct the preset mapping relationship and the preset warning threshold according to the prediction deviation.

[0012] Periodically aggregate historical data on the frequency of calibration cycle compression commands, cumulative compression magnitude, and changes in the comprehensive degradation trend index into interim correction data, and upload them to the cloud-based metrology data center;

[0013] The system receives the long-term baseline calibration cycle issued by the cloud metering data center and replaces the original local long-term baseline calibration cycle. The long-term baseline calibration cycle is calculated by the cloud using the intermediate correction data to update the degradation curve parameters and environmental stress accumulation model parameters of the same batch of electricity meters.

[0014] By reading six types of self-diagnostic signals, such as the reference voltage monitoring value and the analog-to-digital converter gain error value, from inside the metering chip in real time, and building a closed-loop adjustment mechanism based on a comprehensive degradation trend index in the edge computing module, early signs of chip-level degradation can be detected before metering errors appear, thereby triggering real-time compression or extension of the calibration cycle. The adjustment strategy is self-optimized through a proportional-integral-derivative controller, realizing a fundamental transformation of the electricity meter calibration cycle from open-loop static allocation to closed-loop dynamic adaptation.

[0015] Preferably, the comprehensive degradation trend index is calculated using the following mathematical formula:

[0016]

[0017] in, denoted as the comprehensive degradation trend index at time t, dimensionless, ranging from 0 to 1. 'i' is the self-diagnostic signal category index, with values ​​from 1 to 6, corresponding to the reference voltage monitoring signal, analog-to-digital converter gain error signal, temperature sensor signal, power supply voltage monitoring signal, phase error detection signal, and clock frequency deviation signal, respectively. Let be the dynamic weighting coefficient of the i-th type of self-diagnostic signal at time t, satisfying that the sum of the weighting coefficients of all six types of self-diagnostic signals is 1. The value is determined by referring to a table based on the cumulative operating years of the electricity meter. Let be the real-time sampled value of the i-th type of self-diagnostic signal at time t. The baseline value of the i-th type of self-diagnostic signal is taken from the first thirty consecutive days after the initial installation and operation of the electricity meter. The arithmetic mean, The maximum permissible value for the i-th type of self-diagnostic signal is taken as the upper limit of the permissible value for this type of signal specified in the electricity meter's manufacturer's technical specifications.

[0018] By employing the aforementioned mathematical formula to weight and fuse six types of self-diagnostic signals to calculate the comprehensive degradation trend index, the weight coefficients of each signal category can be dynamically adjusted according to the cumulative operating years of the electricity meter. This ensures that the assessment of degradation trends always matches the actual sensitivity of each type of signal at different operating stages, thereby improving the accuracy of the comprehensive degradation trend index and the reliability of calibration cycle adjustment decisions.

[0019] Preferably, the six types of self-diagnostic signals are:

[0020] The reference voltage deviation value output by the reference voltage monitoring register is used to reflect the degree of drift of the internal reference voltage source of the metering chip relative to the factory calibration value;

[0021] The gain error value output by the analog-to-digital converter gain error register is used to reflect the degree of deviation between the actual gain of the analog-to-digital conversion module inside the metering chip and the calibration gain.

[0022] The chip temperature value output by the temperature sensor is used to reflect the operating temperature environment of the metering chip;

[0023] The power supply voltage deviation value output by the power supply voltage monitoring circuit is used to reflect the degree of deviation of the metering chip's supply voltage from the nominal value;

[0024] The phase error value output by the phase error detection register is used to reflect the change in phase difference between the voltage sampling channel and the current sampling channel of the metering chip;

[0025] The clock frequency deviation value output by the clock frequency deviation register is used to reflect the degree of frequency drift of the crystal oscillator inside the metering chip.

[0026] By incorporating the reference voltage deviation, analog-to-digital converter gain error, chip temperature, power supply voltage deviation, phase error, and clock frequency deviation into the self-diagnostic signal sensing system, the performance degradation process of the metering chip can be comprehensively captured from multiple physical dimensions. This avoids misjudgment of the degree of degradation due to incomplete or interfered information from a single signal source, providing a multi-source redundant sensing basis for the accurate adjustment of the calibration cycle.

[0027] Preferably, the edge computing module adopts a three-layer time-scale coupled calibration cycle hierarchical adjustment architecture:

[0028] The first layer is the long-term scale, which is output by the cloud metering data center based on the electricity meter manufacturing process data, the historical degradation curves of the same batch, and the environmental stress accumulation model to determine the long-term baseline calibration cycle.

[0029] The second layer is the medium-term scale, where the edge computing module performs a medium-term correction on the long-term baseline calibration cycle based on the remote error online monitoring results and the medium-term correction data, and outputs a recommended calibration cycle range.

[0030] The third layer is a short-term scale, in which the edge computing module fine-tunes the recommended calibration cycle range in real time based on calibration cycle compression instructions and calibration cycle extension instructions;

[0031] The coupling mechanism among the three layers is as follows: the mid-term correction results serve as the correction factor input for the environmental stress accumulation model in the next long-term scale, while the compression frequency and magnitude of the short-term scale inversely affect the confidence weight of the mid-term scale state assessment.

[0032] By constructing a three-layer time-coupled calibration cycle hierarchical adjustment architecture with long-term, medium-term, and short-term scales, and enabling mutual feedback and correction of decision results among the three layers, it can simultaneously take into account the smooth prediction of long-term degradation trends, the dynamic correction of medium-term state assessments, and the rapid response to short-term abnormal events, effectively avoiding the shortcomings of single-time-scale adjustment architectures that suffer from neglecting one aspect for another.

[0033] Preferably, in the three-layer time-scale coupled calibration cycle hierarchical adjustment architecture, the weight coefficients of the three time scales during weighted fusion are dynamically adjusted according to the operating years of the electricity meter. The operating years of the electricity meter are calculated from the date of its first installation and power-on operation. For each additional year of operating years, the weight coefficient of the long-term scale decreases by a first preset step, the weight coefficient of the short-term scale increases by a second preset step, and the weight coefficient of the medium-term scale remains at 1 minus the sum of the weight coefficients of the long-term and short-term scales.

[0034] By dynamically adjusting the weighting coefficients of the three time scales in the fusion decision based on the operating years of the electricity meter, the calibration cycle adjustment strategy can be adapted to the life stage of the electricity meter. In the early stage of operation, long-term degradation prediction is the main focus, while in the later stage of operation, the response weight to short-term real-time signals is gradually increased, thereby improving the rationality and pertinence of the calibration cycle adjustment throughout the entire life cycle.

[0035] Preferably, the preset mapping relationship is a piecewise linear mapping function, with the rate of change of the comprehensive degradation trend index as the input variable and the compression coefficient as the output variable. The compression coefficient is defined as the ratio of the remaining calibration period after compression to the remaining calibration period before compression.

[0036] When the rate of change of the comprehensive degradation trend index is greater than the preset warning threshold, the compression coefficient decreases linearly with the increase of the rate of change, and the minimum compression coefficient is not lower than the preset lower limit.

[0037] When the rate of change of the comprehensive degradation trend index is lower than the preset warning threshold but higher than the preset attention threshold, the compression coefficient remains unchanged.

[0038] When the absolute value of the rate of change of the comprehensive degradation trend index is lower than the preset stable threshold in multiple consecutive monitoring windows, the extension coefficient takes a preset fixed value, and the remaining calibration period after extension does not exceed the long-term baseline calibration period multiplied by a preset upper limit multiple.

[0039] By using a piecewise linear mapping function with the rate of change of the comprehensive degradation trend index as input to generate the calibration cycle compression coefficient or extension coefficient, the cycle adjustment amplitude that matches the detected degradation acceleration signal of different levels can be output. This avoids overreaction to slight degradation signals and ensures timely intervention to rapid degradation signals, so that the cycle adjustment intensity and the severity of degradation maintain a quantitative correspondence.

[0040] Preferably, the proportional term of the incremental proportional-integral-derivative controller adjusts the slope of the compression coefficient mapping curve according to the current prediction deviation, the integral term accumulates the sum of historical prediction deviations to eliminate systematic prediction deviations, and the derivative term suppresses the oscillations of the calibration cycle adjustment based on the changing trend of the prediction deviation.

[0041] The output of the incremental proportional-integral-derivative controller is a correction increment. It is added to the current parameter with a preset mapping relationship to obtain the updated mapping parameter. The updated mapping parameter is used to generate subsequent calibration cycle compression instructions.

[0042] By using an incremental proportional-integral-derivative controller to correct the preset mapping relationship using proportional, integral, and derivative terms respectively, the mapping parameters can be continuously optimized based on the measured error feedback after each periodic adjustment. This eliminates systematic over-adjustment or under-adjustment deviations and suppresses periodic oscillations caused by improper adjustment parameters, thereby continuously improving the decision accuracy of the calibration cycle as the system runs.

[0043] Preferably, the interim revision data includes:

[0044] The total number of times calibration cycle compression commands and calibration cycle extension commands were triggered in the past ninety days;

[0045] The product of the compression coefficients in all compression commands over the past ninety days;

[0046] Statistical distribution of the categories of reasons for each compression command trigger;

[0047] The changing trend of the monthly average value of the comprehensive degradation trend index.

[0048] By incorporating the total number of triggers, the product of compression coefficients, the distribution of trigger cause categories, and the monthly trend of the comprehensive degradation trend index into the mid-term correction data, the behavioral characteristics accumulated in the short-term closed-loop adjustment can be transferred to the mid-term state evaluation stage. This provides quantifiable behavioral evidence for the mid-term calibration cycle correction and enhances the coupling depth between the short and mid-term in the hierarchical architecture.

[0049] Preferably, the environmental stress accumulation model is constructed based on the Arrhenius equation, which converts the environmental stress actually experienced by the electricity meter into an equivalent operating period;

[0050] The environmental stress includes the cumulative number of high-temperature hours, the cumulative number of high-humidity hours, the cumulative number of temperature cycles, and the cumulative total harmonic distortion.

[0051] The equivalent operating years are used to adjust the predicted degradation rate of the energy meter in the degradation curve of the same batch. The longer the equivalent operating years, the shorter the long-term baseline calibration cycle.

[0052] By using an environmental stress accumulation model based on the Arrhenius equation to convert the cumulative high-temperature hours, cumulative high-humidity hours, cumulative temperature cycles, and cumulative total harmonic distortion experienced by the electricity meter into equivalent operating years, the model can more accurately reflect the accelerated aging effect of complex field environments on the internal components of the electricity meter. This makes the calculation of the long-term baseline calibration cycle closer to the actual health status of the electricity meter.

[0053] Secondly, a dynamic adjustment system for the calibration cycle of electricity meters, applicable to a dynamic adjustment method for the calibration cycle of electricity meters, comprises:

[0054] The end-layer data acquisition unit is deployed inside the metering chip of each energy meter. It is used to read six types of self-diagnostic signals in real time from the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register and clock frequency deviation register inside the metering chip.

[0055] The edge computing unit, deployed in the area concentrator or edge computing gateway, includes a self-diagnostic signal fusion and degradation trend analysis subunit, a short-term calibration cycle real-time fine-tuning decision subunit, a proportional-integral-differential parameter self-optimization subunit, and a mid-term state evaluation correction subunit.

[0056] The cloud-based metering data unit, deployed in the provincial metering data center, includes a batch degradation curve modeling subunit, an environmental stress accumulation model subunit, a long-term baseline periodic update subunit, and a cross-regional collaborative scheduling subunit.

[0057] The terminal data acquisition unit, the edge computing unit, and the cloud metering data unit are connected sequentially through a wireless communication network, together forming the system hardware architecture to which the method applies.

[0058] The three-layer system hardware architecture, consisting of end-layer data acquisition unit, edge computing unit, and cloud metering data unit, can organically integrate real-time acquisition of chip-level self-diagnostic signals, short- and medium-term closed-loop decision-making at the edge, and long-term degradation modeling and cross-regional collaborative scheduling in the cloud. This provides complete system-level support for achieving a full-link closed loop from perception and decision-making to verification and optimization in the calibration cycle.

[0059] The present invention has the following effects:

[0060] I. This invention directly reads multi-dimensional self-diagnostic signals from the internal metering chip of an energy meter and constructs a real-time adjustment mechanism for the calibration cycle based on a comprehensive degradation trend index. This mechanism can accurately capture early performance degradation signs of the metering chip before metering errors appear. It can trigger dynamic compression or extension of the calibration cycle without waiting for the fixed calibration cycle to end. At the same time, it establishes a complete closed-loop link from signal perception, cycle adjustment to actual measurement feedback and strategy optimization. This solves the defects of the existing open-loop static adjustment mode, such as lag in response and lack of effective perception of early chip-level degradation, and significantly improves the timeliness and reliability of energy meter calibration cycle adjustment.

[0061] Second, this invention constructs a three-tiered calibration cycle adjustment architecture with coupled long-term, medium-term, and short-term time scales, coupled with an edge-cloud collaborative system deployment mode. This architecture can simultaneously take into account the prediction of long-term degradation trends of electricity meters, the correction of medium-term operating status, and the rapid response to short-term abnormal events. It can also dynamically adapt the decision weights of each level based on the entire life cycle of the electricity meter and optimize the long-term baseline cycle by combining the degradation patterns of electricity meters in the same batch with the cumulative effect of environmental stress. Thus, while ensuring metering accuracy, it achieves differentiated and refined management of the electricity meter calibration cycle, effectively reducing the overall cost of power grid metering operation and maintenance and unnecessary resource waste.

[0062] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0063] The following sections will describe some specific embodiments of the invention in a detailed manner by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings:

[0064] Figure 1 This is a schematic diagram of the overall process of the dynamic adjustment method for the calibration cycle of an energy meter according to the present invention;

[0065] Figure 2 This is a schematic diagram of the three-layer time-scale coupled calibration cycle hierarchical adjustment architecture of the present invention.

[0066] Figure 3 This is a schematic diagram of the architecture of the dynamic adjustment system for the calibration cycle of the electricity meter according to the present invention. Detailed Implementation

[0067] The following reference Figures 1 to 3This invention describes a method and system for dynamically adjusting the calibration cycle of an electricity meter, according to embodiments of the present invention. In this description, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature, that is, include one or more of that feature. In the description of the present invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified. When a feature "includes or contains" one or more of the features it encompasses, unless otherwise specifically described, this indicates that other features are not excluded and may be further included.

[0068] In the description of this embodiment, the terms "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0069] Example

[0070] This embodiment provides a complete implementation of a method and system for dynamically adjusting the calibration cycle of electricity meters. This embodiment is applied to the full life cycle operation and maintenance scenario of smart electricity meters in low-voltage distribution network areas. It is compatible with the general technical specifications for smart electricity meters issued by the power industry and can be directly deployed in the existing hardware architecture of existing smart electricity meters, distribution area concentrators, edge computing gateways and provincial metering data centers without the need for large-scale hardware modifications to existing equipment.

[0071] like Figure 3 As shown, the system's three-layer hardware architecture and internal sub-unit division, as well as the communication connections between each unit, are fully demonstrated. The system in this embodiment includes an end-layer data acquisition unit, an edge computing unit, and a cloud-based metering data unit. These three types of units are connected sequentially through a wireless communication network, forming a complete hardware support system.

[0072] The end-level data acquisition unit is deployed inside the metering chip of each energy meter. The metering chip adopts a dedicated metering chip for three-phase smart energy meters, which integrates a complete metering sampling channel, a reference voltage source, an analog-to-digital converter module, a temperature sensor, a power supply voltage monitoring circuit, a phase error detection module, and a clock frequency monitoring module. It can directly output various self-diagnostic signals. The core function of the end-level data acquisition unit is to read six types of self-diagnostic signals in real time from the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register, and clock frequency deviation register inside the metering chip.

[0073] Specifically, the reference voltage monitoring register is a dedicated 16-bit read-only register inside the metering chip. The register output value is the deviation between the real-time sampled value of the reference voltage and the factory calibration value. The sampling resolution is 100 microvolts per word, and the reading cycle is 1 minute per read. The output reference voltage deviation value directly reflects the degree of drift of the internal reference voltage source of the metering chip relative to the factory calibration value.

[0074] The analog-to-digital converter gain error register is a dedicated 16-bit read-only register inside the metering chip. The register output value is the deviation between the real-time gain of the analog-to-digital converter module and the calibration gain. The sampling resolution is 0.001% per word, and the reading cycle is 1 minute per read. The output gain error value directly reflects the degree of deviation between the actual gain of the analog-to-digital converter module inside the metering chip and the calibration gain.

[0075] The temperature sensor is integrated into the core area of ​​the metering chip. It is a bandgap temperature sensor with a measurement range of -40°C to 125°C, a measurement resolution of 0.5°C, and a reading cycle of 1 minute. The output chip temperature value directly reflects the operating temperature environment of the metering chip.

[0076] The power supply voltage monitoring circuit is integrated inside the metering chip. It is a high-precision voltage divider sampling circuit with a sampling range of 2.7V to 5.5V, a sampling resolution of 1mV per digit, and a reading cycle of 1 minute. The output power supply voltage deviation value directly reflects the degree of deviation of the metering chip's power supply voltage from the nominal value.

[0077] The phase error detection register is a dedicated 16-bit read-only register inside the metering chip. The register output value is the deviation between the real-time phase difference and the nominal phase difference between the voltage sampling channel and the current sampling channel. The sampling resolution is 0.001 degrees per word, and the reading cycle is 1 minute per read. The output phase error value directly reflects the change in phase difference between the voltage sampling channel and the current sampling channel of the metering chip.

[0078] The clock frequency deviation register is a dedicated 16-bit read-only register inside the metering chip. The register output value is the deviation between the real-time frequency and the nominal frequency of the internal crystal oscillator. The sampling resolution is 0.1 ppm per word, and the reading cycle is 1 minute per read. The output clock frequency deviation value directly reflects the degree of frequency drift of the internal crystal oscillator of the metering chip.

[0079] The six types of self-diagnostic signals read by the end-layer data acquisition unit are encapsulated according to the power industry standard communication protocol via the power line carrier communication module or low-power wireless communication module built into the energy meter, and then uploaded to the edge computing module deployed in the distribution area concentrator or edge computing gateway. The normal data upload cycle is once every 15 minutes. When the change in the sampled value of any type of self-diagnostic signal exceeds the preset maximum reasonable change threshold between two consecutive samples, the upload cycle is automatically shortened to once every 1 minute to ensure the real-time transmission of abnormal data. The data frame permanently includes the energy meter's unique asset number, data acquisition timestamp, real-time sampled values ​​of the six types of self-diagnostic signals, and data check code to ensure data traceability and accuracy.

[0080] The edge computing unit is deployed in the distribution center concentrator or edge computing gateway. The edge computing gateway uses an ARM Cortex-A53 architecture processor with no less than 1GB of RAM and no less than 8GB of storage, and runs on a Linux operating system to meet the real-time computing and data storage requirements of the edge side. The edge computing unit includes a self-diagnostic signal fusion and degradation trend analysis subunit, a short-term calibration cycle real-time fine-tuning decision subunit, a proportional-integral-differential parameter self-optimization subunit, and a mid-term state evaluation and correction subunit. These four subunits work together in sequence to complete all data processing and decision-making functions at the edge side.

[0081] The cloud-based metering data unit is deployed in the provincial metering data center and adopts a distributed cloud computing architecture. It has massive data storage and parallel computing capabilities, and can process uploaded data from more than 100,000 distribution stations simultaneously. The cloud-based metering data unit includes a batch degradation curve modeling subunit, an environmental stress accumulation model subunit, a long-term baseline periodic update subunit, and a cross-distribution station collaborative scheduling subunit. The four subunits work together to complete the long-term degradation trend modeling and baseline periodic update of electricity meters throughout the region.

[0082] like Figure 1 As shown, this fully demonstrates the entire process chain from edge-side self-diagnostic signal acquisition to edge-side data processing and periodic adjustment decisions, and then to long-term baseline updates in the cloud, as well as the closed-loop feedback optimization path formed by the incremental proportional-integral-derivative controller. The method in this embodiment is executed by the edge computing module and specifically includes the following operations.

[0083] The system receives six types of self-diagnostic signals uploaded by the electricity meter. These six types of self-diagnostic signals are output in real time by the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register, and clock frequency deviation register deployed in the metering chip within the electricity meter.

[0084] Specifically, the edge computing module receives six types of self-diagnostic signals uploaded by the energy meter through the communication module. These six types of self-diagnostic signals are read from the registers and circuits corresponding to the metering chip by the end-layer data acquisition unit according to a preset reading cycle. Each sampled value is initially checked for its range, and abnormal sampled values ​​that exceed the physical range are eliminated. The valid sampled values ​​from the previous moment are used to replace the abnormal values ​​to ensure the basic validity of the sampled data.

[0085] After receiving a data frame, the edge computing module first performs a checksum verification. Once the verification is successful, the data is stored in the local database.

[0086] Specifically, the communication module of the electricity meter encapsulates six types of self-diagnostic signals according to a protocol and sends them to the edge computing module according to a preset upload cycle. After receiving the data frame, the edge computing module first performs a checksum verification. If the verification is successful, the data is stored in the local database. If the verification fails, a data retransmission command is sent to the electricity meter. A maximum of three retransmissions are allowed. If three retransmissions still fail, the data frame is discarded and replaced with the previous valid data frame to ensure the accuracy and continuity of data transmission.

[0087] After judging the validity of the data from the six types of self-diagnostic signals, the edge computing module calculates the comprehensive degradation trend index and the rate of change of the comprehensive degradation trend index within a preset sliding time window.

[0088] Specifically, the self-diagnostic signal fusion and degradation trend analysis subunit of the edge computing module first performs a secondary data validity judgment on the received six types of self-diagnostic signals. The validity judgment rules include three items: first, whether the sampled value is within the allowable range specified in the manufacturer's specifications for that type of signal; second, whether the sampling timestamp is continuous with the system standard time; and third, whether the change in the sampled value of the same signal between two consecutive samples exceeds a preset maximum reasonable change threshold. Sampled values ​​that meet all three rules are determined to be valid data, while sampled values ​​that do not meet any one rule are determined to be invalid data and are directly discarded, and the valid sampled value of the signal at the previous moment is used as a replacement.

[0089] After verifying the validity of the data, the self-diagnostic signal fusion and degradation trend analysis subunit calculates the comprehensive degradation trend index. The comprehensive degradation trend index is calculated using the following mathematical formula:

[0090]

[0091] in, is the comprehensive degradation trend index at time t. This parameter is dimensionless and ranges from 0 to 1. The larger the value, the more severe the comprehensive degradation of the electricity meter's metering chip. t is the data sampling time in seconds, using a standard timestamp synchronized with the electricity meter's internal real-time clock. is the summation operator, with an operation range of all integers from i to 6. It represents the summation of the weighted calculation results corresponding to the six types of self-diagnostic signals to obtain the final comprehensive degradation trend index. i is the self-diagnostic signal category index, with values ​​from 1 to 6. Each value corresponds to a fixed type of self-diagnostic signal, where i=1 corresponds to the reference voltage monitoring signal, i=2 corresponds to the analog-to-digital converter gain error signal, i=3 corresponds to the temperature sensor signal, i=4 corresponds to the power supply voltage monitoring signal, i=5 corresponds to the phase error detection signal, and i=6 corresponds to the clock frequency deviation signal. The dynamic weighting coefficient of the i-th type of self-diagnostic signal at time t. The sum of the weighting coefficients of all six types of self-diagnostic signals is fixed at 1. The value of this parameter is determined by a preset weighting table based on the cumulative operating years of the energy meter. The longer the cumulative operating years of the energy meter, the higher the weighting coefficient of the signal that is strongly correlated with chip aging. The sampled value of the i-th type of self-diagnostic signal at time t is the real-time sampled value. The unit of the sampled value is consistent with the physical unit of the corresponding signal. It is directly read by the end-layer data acquisition unit from the register or circuit corresponding to the metering chip. The baseline value of the i-th type of self-diagnostic signal is the arithmetic mean of all valid sampled values ​​of the corresponding type of self-diagnostic signal within 30 consecutive days after the first installation and operation of the energy meter, representing the benchmark performance level of the energy meter in a brand-new state. This is the maximum permissible value for the i-th type of self-diagnostic signal. This value is taken as the upper limit of the permissible value for this type of signal as specified in the technical specifications of the electricity meter manufacturer. It represents the maximum deviation that this type of signal can achieve without affecting the accuracy of the measurement.

[0092] For example, after the electricity meter is first installed and put into operation, the arithmetic mean of the effective sampled values ​​of the reference voltage monitoring signal over thirty consecutive days is 2 microvolts. This value is the value corresponding to when i equals 1. The manufacturer's technical specifications for this electricity meter state that the maximum allowable deviation of the reference voltage is 50 microvolts, which corresponds to the value when i equals 1. The real-time sampled value of the reference voltage monitoring signal at time t. When the value is 12 microvolts, the corresponding absolute deviation is 10 microvolts, and the relative deviation is 10 microvolts divided by 48 microvolts, which is approximately 0.208. If the value at that moment corresponds to... If the value is 0.25, then the contribution of this signal to the comprehensive degradation trend index is 0.25 multiplied by 0.208, which is approximately 0.052. After calculating the contribution values ​​of all six types of signals, the final result is obtained by summing them up. Numerical value.

[0093] After calculating the comprehensive degradation trend index, the self-diagnostic signal fusion and degradation trend analysis subunit calculates the rate of change of the comprehensive degradation trend index within a preset sliding time window. The preset sliding time window is 24 hours long, and the sliding step is 1 hour. The rate of change is calculated by dividing the difference between the average value of the comprehensive degradation trend index in the current sliding time window and the average value of the comprehensive degradation trend index in the previous adjacent sliding time window by the length of the sliding time window, in units of hourly. A positive rate of change indicates an accelerating degradation trend, a negative value indicates a slowing degradation trend, and a value of 0 indicates a stable degradation trend.

[0094] When the rate of change of the comprehensive degradation trend index exceeds the preset warning threshold, the edge computing module automatically generates a calibration cycle compression command without waiting for the current calibration cycle to end. The remaining calibration cycle of the electricity meter is compressed to a set proportion of the original remaining calibration cycle according to the preset mapping relationship. When the absolute value of the rate of change of the comprehensive degradation trend index is lower than the preset stability threshold in multiple consecutive monitoring windows, the module automatically generates a calibration cycle extension command and extends the remaining calibration cycle of the electricity meter according to the preset extension coefficient.

[0095] Specifically, the edge computing module's short-term calibration cycle real-time fine-tuning decision subunit receives data on the rate of change of the comprehensive degradation trend index and executes dynamic adjustment decisions for the calibration cycle. For example... Figure 2 As shown, the input sources and output content of the three levels—long-term, medium-term, and short-term—as well as the coupling and feedback relationships between the three levels, are illustrated. The dynamic adjustment rules of the weight coefficients of the three levels with the operating years of the electricity meter are also demonstrated. The edge computing module adopts a three-level time-scale coupled calibration cycle hierarchical adjustment architecture, which is divided into three levels: long-term, medium-term, and short-term.

[0096] The first layer is the long-term scale, which is output by the cloud metering data center based on the electricity meter manufacturing process data, historical degradation curves of the same batch, and environmental stress accumulation model. The long-term baseline calibration cycle is the benchmark value of the electricity meter calibration cycle, representing the recommended calibration cycle of the same batch of electricity meters under standard operating conditions.

[0097] The second layer is the medium-scale, where the edge computing module performs medium-scale corrections on the long-term baseline calibration cycle based on the results of remote error online monitoring and medium-scale correction data, and outputs a recommended calibration cycle range. The recommended calibration cycle range includes the upper and lower limits of the calibration cycle, providing a constraint range for real-time fine-tuning at the short-term scale.

[0098] The third layer is the short-term scale, where the edge computing module performs real-time fine-tuning of the recommended calibration cycle range based on calibration cycle compression instructions and calibration cycle extension instructions, and outputs the final effective calibration cycle.

[0099] The coupling mechanism among the three layers is as follows: the mid-term correction results serve as the correction factor input for the environmental stress accumulation model in the next long-term timescale, while the compression frequency and magnitude of the short-term timescale inversely affect the confidence weight of the mid-term timescale state assessment. The weighting coefficients of the three timescales during weighted fusion are dynamically adjusted based on the operating years of the electricity meter. The operating years are calculated from the date the meter is first installed and powered on. For each additional year, the long-term timescale weighting coefficient decreases by a first preset step, the short-term timescale weighting coefficient increases by a second preset step, and the mid-term timescale weighting coefficient remains at 1 minus the sum of the long-term and short-term timescale weighting coefficients. The first preset step is set to 0.05, the second preset step is set to 0.04, and upper and lower limits are set for the adjustment range of the weighting coefficients. The minimum long-term timescale weighting coefficient is no less than 0.2, and the maximum short-term timescale weighting coefficient is no more than 0.5, ensuring the stability of the three-layer architecture.

[0100] For example, in the first year of operation of an electricity meter, the long-term weighting factor is set to 0.6, the short-term weighting factor to 0.1, and the medium-term weighting factor to 0.3. In the fifth year of operation, the long-term weighting factor decreases to 0.4, the short-term weighting factor increases to 0.3, and the medium-term weighting factor remains at 0.3. In the tenth year of operation, the long-term weighting factor decreases to 0.2, the short-term weighting factor increases to 0.5, and the medium-term weighting factor remains at 0.3.

[0101] Specifically, the preset mapping relationship is a piecewise linear mapping function, with the rate of change of the comprehensive degradation trend index as the input variable and the compression coefficient as the output variable. The compression coefficient is defined as the ratio of the remaining calibration period after compression to the remaining calibration period before compression. The preset warning threshold is set to 0.005 per hour, the preset attention threshold is set to 0.001 per hour, the preset stability threshold is set to 0.0005 per hour, the preset lower limit compression coefficient is set to 0.2, the preset upper limit multiple is set to 1.5, and the preset extension coefficient is a fixed value of 1.1.

[0102] The piecewise linear mapping function is divided into three intervals. In the first interval, when the rate of change of the comprehensive degradation trend index exceeds a preset warning threshold, the compression coefficient decreases linearly with increasing rate of change, and the minimum compression coefficient is not lower than a preset lower limit. In the second interval, when the rate of change of the comprehensive degradation trend index is lower than a preset warning threshold but higher than a preset attention threshold, the compression coefficient remains unchanged, and no compression operation is performed on the calibration cycle. In the third interval, when the absolute value of the rate of change of the comprehensive degradation trend index is lower than a preset stability threshold for multiple consecutive monitoring windows, the extension coefficient takes a preset fixed value, and the remaining calibration cycle after extension does not exceed the long-term baseline calibration cycle multiplied by a preset upper limit multiple. The number of consecutive monitoring windows is set to three.

[0103] For example, when the rate of change of the comprehensive degradation trend index is 0.006 per hour, exceeding the preset warning threshold of 0.005 per hour, the calculated compression coefficient is 0.8. If the original remaining calibration period of the electricity meter is 12 months, the compressed remaining calibration period is 9.6 months, which takes effect immediately without waiting for the current calibration period to end. When the rate of change of the comprehensive degradation trend index is 0.003 per hour, falling between the preset attention threshold and the preset warning threshold, the compression coefficient remains at 1, and no calibration period adjustment is performed. If the absolute value of the rate of change of the comprehensive degradation trend index is less than 0.0005 per hour for three consecutive sliding time windows, the extension coefficient is 1.1. If the original remaining calibration period of the electricity meter is 10 months, the extended remaining calibration period is 11 months. If the long-term baseline calibration period is 12 months and the preset upper limit multiple is 1.5, the maximum allowable extended remaining calibration period is 18 months. Since 11 months does not exceed the upper limit, the adjustment instruction takes effect immediately.

[0104] The edge computing module compares the compression range of the calibration cycle compression command and the extension range of the calibration cycle extension command with the measured error value obtained from the subsequent actual calibration of the energy meter, calculates the predicted deviation, and uses an incremental proportional-integral-derivative controller to correct the preset mapping relationship and preset warning threshold based on the predicted deviation.

[0105] Specifically, the proportional-integral-derivative (PID) parameter self-optimization subunit of the edge computing module receives the measured error value obtained from the on-site calibration of the power meter and performs closed-loop self-optimization of the parameters. First, it calculates the prediction deviation, which is the difference between the calibration cycle adjustment range corresponding to the measured error and the adjustment range generated by the current instruction. A positive difference indicates insufficient adjustment, while a negative difference indicates excessive adjustment. The incremental PID controller takes the prediction deviation as input and outputs the correction increment. The correction increment is accumulated with the current parameters of the preset mapping relationship to obtain the updated mapping parameters. These updated mapping parameters are used to generate subsequent calibration cycle compression instructions.

[0106] The incremental proportional-integral-derivative (PID) controller consists of three parts: a proportional term, an integral term, and a derivative term. These three parts work together to correct parameters. The proportional term adjusts the slope of the compression coefficient mapping curve based on the magnitude of the current prediction deviation. The proportional coefficient is set to 0.8; the larger the current prediction deviation, the greater the adjustment of the mapping curve slope. The integral term accumulates the sum of historical prediction deviations to eliminate systematic prediction errors. The integral coefficient is set to 0.1, continuously accumulating historical deviations to eliminate systematic errors in a fixed direction. The derivative term suppresses oscillations in the calibration cycle adjustment based on the trend of prediction deviation changes. The derivative coefficient is set to 0.05; when the prediction deviation fluctuates rapidly, it automatically suppresses the adjustment amplitude to avoid oscillations in the cycle adjustment.

[0107] For example, if the prediction deviations of five consecutive calibrations are all positive, it indicates that the system has a persistent systematic bias of insufficient adjustment. After the integral term accumulates historical deviations, it automatically increases the slope of the compression coefficient mapping curve, thereby increasing the compression amplitude corresponding to the same rate of change and eliminating the systematic bias. When the prediction deviation shows continuous alternating positive and negative fluctuations, it indicates that the periodic adjustment is oscillating. The derivative term automatically reduces the adjustment amplitude to suppress the oscillation and ensure the stability of the adjustment process.

[0108] The edge computing module periodically aggregates historical data on the frequency of calibration cycle compression commands, cumulative compression magnitude, and changes in the comprehensive degradation trend index into interim correction data, which is then uploaded to the cloud-based metering data center.

[0109] Specifically, the mid-term status evaluation and correction subunit of the edge computing module aggregates and uploads mid-term correction data once per calendar month. The mid-term correction data comprises four parts: The first part is the total number of triggers for calibration cycle compression and extension commands over the past ninety days, respectively, to characterize the fluctuation level of the electricity meter's status. The second part is the product of the compression coefficients in each compression command over the past ninety days, representing the cumulative compression magnitude of the calibration cycle over those ninety days. The third part is the statistical distribution of the triggering reasons for each compression command, identifying the dominant triggering signal category among six self-diagnostic signals for each compression command trigger, used to analyze the main causes of electricity meter degradation. The fourth part is the changing trend of the monthly average of the comprehensive degradation trend index, statistically analyzing the average comprehensive degradation trend index for each calendar month over the past twelve months to form continuous trend data. When uploading mid-term correction data, the unique asset number, operating years, and installation area number of the corresponding electricity meter are included to ensure the traceability of cloud data.

[0110] The edge computing module uploads the interim correction data to the cloud metering data unit. Upon receiving this interim correction data, the cloud metering data unit's batch degradation curve modeling subunit divides electricity meters from the same manufacturer, model, and production batch into batch groups. Using the interim correction data from all electricity meters within the group, it updates the degradation curve parameters of the batch's electricity meters. The environmental stress accumulation model subunit constructs an environmental stress accumulation model based on the Arrhenius equation, converting the actual environmental stress experienced by the electricity meters into equivalent operating years. The long-term baseline periodic update subunit calculates the long-term baseline calibration period for each electricity meter based on the updated degradation curve parameters and equivalent operating years. The edge computing module receives this long-term baseline calibration period from the cloud metering data unit and replaces the existing local long-term baseline calibration period, completing the long-term periodic update.

[0111] Therefore, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications conforming to the principles of the present invention can be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the invention. Thus, the scope of the present invention should be understood and construed as covering all such other variations or modifications.

Claims

1. A method for dynamically adjusting the calibration cycle of an electricity meter, characterized in that, This method is executed by an edge computing module deployed in a distribution center concentrator or edge computing gateway, and includes: The device receives six types of self-diagnostic signals uploaded by the energy meter. These six types of self-diagnostic signals are output in real time by the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register, and clock frequency deviation register inside the energy meter's metering chip. After judging the validity of the six types of self-diagnostic signals received, the comprehensive degradation trend index is calculated, and the rate of change of the comprehensive degradation trend index within the preset sliding time window is calculated. When the rate of change of the comprehensive degradation trend index exceeds the preset warning threshold, the calibration cycle compression command is automatically generated without waiting for the current calibration cycle to end. The remaining calibration cycle of the energy meter is compressed to a set proportion of the original remaining calibration cycle according to the preset mapping relationship. When the absolute value of the rate of change of the comprehensive degradation trend index is lower than the preset stability threshold in multiple consecutive monitoring windows, the calibration cycle extension command is automatically generated to extend the remaining calibration cycle of the energy meter according to the preset extension coefficient. The compression range of the calibration cycle compression command and the extension range of the calibration cycle extension command are compared with the measured error value obtained in the subsequent actual calibration of the energy meter to calculate the prediction deviation. The incremental proportional-integral-derivative controller is used to correct the preset mapping relationship and the preset warning threshold according to the prediction deviation. Periodically aggregate historical data on the frequency of calibration cycle compression commands, cumulative compression magnitude, and changes in the comprehensive degradation trend index into interim correction data, and upload them to the cloud-based metrology data center; The system receives the long-term baseline calibration cycle issued by the cloud metering data center and replaces the original local long-term baseline calibration cycle. The long-term baseline calibration cycle is calculated by the cloud using the intermediate correction data to update the degradation curve parameters and environmental stress accumulation model parameters of the same batch of electricity meters.

2. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The comprehensive degradation trend index is calculated using the following mathematical formula: in, denoted as the comprehensive degradation trend index at time t, dimensionless, ranging from 0 to 1. 'i' is the self-diagnostic signal category index, with values ​​from 1 to 6, corresponding to the reference voltage monitoring signal, analog-to-digital converter gain error signal, temperature sensor signal, power supply voltage monitoring signal, phase error detection signal, and clock frequency deviation signal, respectively. Let be the dynamic weighting coefficient of the i-th type of self-diagnostic signal at time t, satisfying that the sum of the weighting coefficients of all six types of self-diagnostic signals is 1. The value is determined by referring to a table based on the cumulative operating years of the electricity meter. Let be the real-time sampled value of the i-th type of self-diagnostic signal at time t. The baseline value of the i-th type of self-diagnostic signal is taken from the first thirty consecutive days after the initial installation and operation of the electricity meter. The arithmetic mean, The maximum permissible value for the i-th type of self-diagnostic signal is taken as the upper limit of the permissible value for this type of signal specified in the electricity meter's manufacturer's technical specifications.

3. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The six types of self-diagnostic signals are as follows: The reference voltage deviation value output by the reference voltage monitoring register is used to reflect the degree of drift of the internal reference voltage source of the metering chip relative to the factory calibration value; The gain error value output by the analog-to-digital converter gain error register is used to reflect the degree of deviation between the actual gain of the analog-to-digital conversion module inside the metering chip and the calibration gain. The chip temperature value output by the temperature sensor is used to reflect the operating temperature environment of the metering chip; The power supply voltage deviation value output by the power supply voltage monitoring circuit is used to reflect the degree of deviation of the metering chip's supply voltage from the nominal value; The phase error value output by the phase error detection register is used to reflect the change in phase difference between the voltage sampling channel and the current sampling channel of the metering chip; The clock frequency deviation value output by the clock frequency deviation register is used to reflect the degree of frequency drift of the crystal oscillator inside the metering chip.

4. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The edge computing module adopts a three-layer time-scale coupled calibration cycle hierarchical adjustment architecture: The first layer is the long-term scale, which is output by the cloud metering data center based on the electricity meter manufacturing process data, the historical degradation curves of the same batch, and the environmental stress accumulation model to determine the long-term baseline calibration cycle. The second layer is the medium-term scale, where the edge computing module performs a medium-term correction on the long-term baseline calibration cycle based on the remote error online monitoring results and the medium-term correction data, and outputs a recommended calibration cycle range. The third layer is a short-term scale, in which the edge computing module fine-tunes the recommended calibration cycle range in real time based on calibration cycle compression instructions and calibration cycle extension instructions; The coupling mechanism among the three layers is as follows: the mid-term correction results serve as the correction factor input for the environmental stress accumulation model in the next long-term scale, while the compression frequency and magnitude of the short-term scale inversely affect the confidence weight of the mid-term scale state assessment.

5. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 4, characterized in that, In the three-layer time-scale coupled calibration cycle hierarchical adjustment architecture, the weight coefficients of the three time scales during weighted fusion are dynamically adjusted according to the operating years of the electricity meter. The operating years of the electricity meter are calculated from the date of its first installation and power-on operation. For each additional year of operating years, the weight coefficient of the long-term scale decreases by a first preset step, the weight coefficient of the short-term scale increases by a second preset step, and the weight coefficient of the medium-term scale remains at 1 minus the sum of the weight coefficients of the long-term and short-term scales.

6. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The preset mapping relationship is a piecewise linear mapping function, with the rate of change of the comprehensive degradation trend index as the input variable and the compression coefficient as the output variable. The compression coefficient is defined as the ratio of the remaining calibration period after compression to the remaining calibration period before compression. When the rate of change of the comprehensive degradation trend index is greater than the preset warning threshold, the compression coefficient decreases linearly with the increase of the rate of change, and the minimum compression coefficient is not lower than the preset lower limit. When the rate of change of the comprehensive degradation trend index is lower than the preset warning threshold but higher than the preset attention threshold, the compression coefficient remains unchanged. When the absolute value of the rate of change of the comprehensive degradation trend index is lower than the preset stable threshold in multiple consecutive monitoring windows, the extension coefficient takes a preset fixed value, and the remaining calibration period after extension does not exceed the long-term baseline calibration period multiplied by a preset upper limit multiple.

7. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The proportional term of the incremental proportional-integral-derivative controller adjusts the slope of the compression coefficient mapping curve according to the current prediction deviation, the integral term accumulates the sum of historical prediction deviations to eliminate systematic prediction deviations, and the derivative term suppresses the oscillations of calibration cycle adjustment based on the changing trend of prediction deviations. The output of the incremental proportional-integral-derivative controller is a correction increment. It is added to the current parameter with a preset mapping relationship to obtain the updated mapping parameter. The updated mapping parameter is used to generate subsequent calibration cycle compression instructions.

8. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The interim revision data includes: The total number of times calibration cycle compression commands and calibration cycle extension commands were triggered in the past ninety days; The product of the compression coefficients in all compression commands over the past ninety days; Statistical distribution of the categories of reasons for each compression command trigger; The changing trend of the monthly average value of the comprehensive degradation trend index.

9. The method for dynamically adjusting the calibration cycle of an electricity meter according to claim 1, characterized in that, The environmental stress accumulation model is based on the Arrhenius equation and converts the environmental stress actually experienced by the electricity meter into an equivalent operating period. The environmental stress includes the cumulative number of high-temperature hours, the cumulative number of high-humidity hours, the cumulative number of temperature cycles, and the cumulative total harmonic distortion. The equivalent operating years are used to adjust the predicted degradation rate of the energy meter in the degradation curve of the same batch. The longer the equivalent operating years, the shorter the long-term baseline calibration cycle.

10. The dynamic adjustment system for calibration cycle of an electricity meter according to claim 1, applicable to the dynamic adjustment method for calibration cycle of an electricity meter according to any one of claims 1 to 9, characterized in that, The system consists of: The end-layer data acquisition unit is deployed inside the metering chip of each energy meter. It is used to read six types of self-diagnostic signals in real time from the reference voltage monitoring register, analog-to-digital converter gain error register, temperature sensor, power supply voltage monitoring circuit, phase error detection register and clock frequency deviation register inside the metering chip. The edge computing unit, deployed in the area concentrator or edge computing gateway, includes a self-diagnostic signal fusion and degradation trend analysis subunit, a short-term calibration cycle real-time fine-tuning decision subunit, a proportional-integral-differential parameter self-optimization subunit, and a mid-term state evaluation correction subunit. The cloud-based metering data unit, deployed in the provincial metering data center, includes a batch degradation curve modeling subunit, an environmental stress accumulation model subunit, a long-term baseline periodic update subunit, and a cross-regional collaborative scheduling subunit.

Images