Dynamic adaptive operation, maintenance and metering optimization method for electric energy metering box

Abstract

The invention discloses a dynamic adaptive operation, maintenance and metering optimization method for an electric energy metering box, and the method comprises the steps: collecting electrical parameters, environmental parameters and box body state parameters in real time, constructing a three-dimensional data set through preprocessing, and carrying out the local storage and preliminary analysis through edge calculation; the condensation occurrence probability is calculated based on the environmental parameters, and self-adaptive regulation and control are triggered and a ventilation structure is linked; abnormal features are extracted, abnormity is recognized in a rule judgment and trend analysis double-engine mode, and graded early warning is carried out; performing comprehensive state scoring in combination with the three-dimensional data set, and matching a differentiated inspection mechanism; presetting a metering error reference value of a temperature and humidity interval, calculating a compensation value in real time and remotely correcting parameters of the electric energy meter; and establishing a shared data platform of the power supply and the power consumption, and realizing full-process data tracing and automatic report generation. According to the method, the condensation prevention and control accuracy is improved, the abnormal false alarm rate is reduced, the operation and maintenance resource allocation is optimized, the metering accuracy is improved, and the data collaboration and tracing capability is enhanced.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to a dynamic adaptive operation and maintenance and metering optimization method for electricity metering boxes. Background Technology

[0002] As the core equipment for electricity metering in power systems, electricity metering boxes are widely used in various power supply and consumption scenarios, including residential electricity consumption and industrial production. Their operational stability and metering accuracy directly affect the legitimate rights and interests of both power suppliers and consumers, as well as the safe and efficient operation of the power system. With the development of power Internet of Things (IoT) technology, existing electricity metering boxes have gradually been equipped with basic sensing devices and data transmission channels. However, their operation, maintenance, and metering optimization still rely on traditional technical models, resulting in many technical shortcomings that urgently need to be addressed.

[0003] In terms of condensation control, existing technologies mostly employ passive protection methods, such as built-in desiccants, fixed ventilation grilles, or timed start-stop fans, without establishing a dynamic correlation with environmental parameters. Because the temperature and humidity variation patterns differ significantly across installation scenarios, traditional fixed methods cannot accurately predict the risk of condensation, resulting in poor condensation control. This leads to problems such as corrosion of wiring terminals inside the enclosure, decreased insulation performance, and even increased metering errors and short-circuit hazards. Furthermore, desiccants require regular manual replacement, and the fixed start-stop mode of ventilation fans easily leads to energy waste or untimely control.

[0004] In terms of anomaly identification and early warning, existing methods mostly rely on threshold judgment of a single electrical parameter to achieve anomaly monitoring, lacking integrated analysis of the physical state of the enclosure and electrical parameters. This single-dimensional monitoring mode is susceptible to power grid fluctuations and environmental interference, resulting in high false alarm and false negative rates for anomalies such as electricity theft, line overload, poor contact, and aging enclosure seals. Furthermore, anomaly identification largely depends on post-event analysis of remote meter reading data, lacking real-time predictive capabilities. After anomalies occur, early warning information cannot be pushed out in a timely manner, leading to delayed fault handling and increased power supply losses.

[0005] In terms of operation and maintenance scheduling, the current operation and maintenance of electricity metering boxes generally adopts a regular full-area inspection mode. Regardless of the actual operating status of the equipment, a comprehensive inspection is carried out according to a fixed cycle. This mode has blindness: on the one hand, potential hidden dangers may be missed for equipment in a sub-healthy state due to the excessively long inspection cycle; on the other hand, over-inspection of equipment in a stable operating state results in a waste of human and material resources. Moreover, the inspection results rely on the experience judgment of operation and maintenance personnel, which is prone to missed inspections and misjudgments, making it difficult to guarantee the efficiency and accuracy of operation and maintenance.

[0006] Regarding the optimization of metering accuracy, existing technologies lack effective means to address metering errors caused by changes in temperature and humidity. The metering accuracy of core components such as current transformers and electricity meters in electricity metering boxes is easily affected by temperature and humidity fluctuations. Traditional metering error calibration adopts an offline overall calibration mode, with calibration cycles typically ranging from six months to one year. This approach cannot capture error drift caused by environmental changes in real time. Furthermore, the calibration process does not distinguish the source of error, lacks specificity, and lacks real-time verification and dynamic compensation mechanisms after calibration, making it easy for errors to accumulate again and affecting the accuracy of metering data.

[0007] In terms of data collaboration and traceability, the data analysis of existing power metering boxes is mostly limited to a single dimension, only collecting power metering data without integrating environmental parameters, box status parameters and operation and maintenance records. It is impossible to establish a correlation model between environment, status and metering accuracy. The data utilization rate is low and there are barriers to data interaction between power supply and consumption parties. Information transmission such as abnormal warnings, equipment hidden dangers and calibration records is delayed, resulting in poor fault handling and lack of full-process data traceability capabilities, which is not conducive to problem tracing and responsibility identification.

[0008] In addition, existing optimization solutions for electricity metering boxes mostly rely on modifications to the box structure, which not only increases equipment modification costs and compatibility risks, but also fails to fundamentally solve the dynamic adaptation problem between operation and maintenance and metering, making it difficult to balance practicality, accuracy and economy. Summary of the Invention

[0009] The purpose of this invention is to provide a dynamic adaptive operation and maintenance and metering optimization method for electricity metering boxes to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a dynamic adaptive operation and maintenance and metering optimization method for an electricity metering box, comprising the following steps: relying on the existing sensing devices and data transmission channels of the electricity metering box, collecting electrical parameters, environmental parameters and box status parameters in real time, performing noise reduction and normalization preprocessing on the raw data, constructing an electrical-environment-status three-dimensional dataset, completing local storage and preliminary analysis through an edge computing module, and reducing remote transmission redundancy; Based on the pre-processed environmental parameters, the probability of condensation is calculated by a preset prediction model. Combined with the difference between the environmental parameters inside and outside the box, adaptive control logic is triggered to link the existing ventilation structure to achieve condensation prevention and control. Simultaneously, control data is recorded to iteratively optimize model parameters. Anomalies are extracted from electrical parameters and enclosure status parameters, multi-dimensional feature vectors are constructed, and anomalies are identified using a dual-engine mode of rule judgment and trend analysis. Warning levels are divided according to severity, and warning information is generated and pushed to the operation and maintenance terminal and linked to historical data comparison to improve the accuracy of identification. By combining the 3D dataset, the metering box is comprehensively scored, the equipment status level is divided according to the score, and the corresponding inspection cycle and key points are matched to form a closed-loop operation and maintenance mechanism of scoring-scheduling-inspection-re-evaluation. The metering error benchmark values ​​for different temperature and humidity ranges are preset. The edge computing module compares the temperature and humidity inside the box with the benchmark range in real time, calculates the error compensation value, and corrects the metering parameters of the electricity meter through remote commands. The compensation accuracy is verified regularly to ensure metering accuracy. Establish a data sharing platform for both power supply and consumption parties to synchronize early warning information, inspection records, error compensation data, and control logs, enabling full-process data traceability and automatically generating operation and maintenance analysis reports to provide data support for strategy iteration.

[0011] Preferably, the electrical parameters include voltage, current, power factor and power data; the environmental parameters include internal temperature and humidity, external temperature and humidity and atmospheric humidity; and the enclosure status parameters include door magnetic switch status, vibration parameters and humidity difference at the sealing point. Data acquisition adopts a combination of timed and triggered modes. Electrical parameters and environmental parameters are acquired at timed intervals, while door magnetic switch status and humidity difference at the sealing point are acquired at triggered intervals. Vibration parameters are acquired through a combination of timed and triggered modes. The preprocessing process also includes removing sensor interference, filling in missing data, filtering out outliers, and verifying the data validity through a preset data validity threshold to ensure the integrity and reliability of the 3D dataset.

[0012] Preferably, the preset prediction model is based on the dew point temperature calculation principle, and a multi-parameter correlation model is constructed by combining the temperature inside the box and the humidity difference between inside and outside the box. It is iteratively trained by combining historical condensation data, corresponding environmental parameters and control effects to adapt to the climate characteristics and installation scenarios of different regions. The adaptive control logic uses the probability of condensation and the difference between parameters inside and outside the chamber as the core triggering conditions. When the probability reaches the preset threshold and the temperature and humidity difference meets the control requirements, the ventilation structure is activated and the running time is dynamically adjusted according to the size of the difference. When there is no significant difference in parameters between the inside and outside of the chamber but the humidity inside the chamber exceeds the standard, an intermittent ventilation mode is adopted, and the ventilation structure is started and stopped in a cycle at preset intervals to balance the condensation control effect and energy consumption.

[0013] Preferably, the abnormal features include the parameter mutation rate, voltage distortion rate, current fluctuation range and power factor deviation of electrical parameters, as well as the abnormal opening time of the door magnet, vibration mutation amplitude, vibration duration and humidity difference mutation value at the sealing point of the enclosure status parameters. The warning levels are divided into at least Level 1 and Level 2. Level 1 anomalies correspond to serious problems that affect electricity safety and metering accuracy, such as suspected electricity theft, line overload, and short circuit hazards. Level 2 anomalies correspond to potential hazards such as poor line contact, aging enclosure seals, and minor vibration interference. Different levels of early warning information correspond to different push priorities and processing procedures. Level 1 abnormal information is pushed to the operation and maintenance terminal in real time and triggers an audio and visual reminder. Level 2 abnormal information is included in the operation and maintenance to-do list and pushed according to priority.

[0014] Preferably, the comprehensive status score adopts a weighted quantitative evaluation, assigning corresponding weights to the stability of measurement accuracy, environmental adaptability and physical status of the enclosure, with the stability of measurement accuracy having the highest weight, and the other two weights can be dynamically adjusted according to the application scenario. Based on the rating, the equipment is divided into three levels: excellent, qualified, and warning. The inspection cycle of excellent equipment is extended, the inspection cycle of qualified equipment is maintained at the regular cycle and the focus is on investigating key warning items, and the inspection cycle of warning equipment is shortened and on-site maintenance is prioritized. After the inspection is completed, the status score is updated based on the on-site testing data to form a closed-loop evaluation system.

[0015] Preferably, the measurement error reference value is preset according to the meter box type classification, covering common types such as single-phase meter, three-phase meter and meter box connected via current transformer, and each type corresponds to the error reference value in different temperature and humidity ranges; The benchmark values ​​are obtained by combining laboratory standard environment simulation tests with calibration based on actual on-site operating data to ensure accuracy and applicability; Error compensation does not change the hardware structure of the electricity meter. It is achieved by adjusting the built-in software metering coefficient through remote commands. After compensation, the accuracy is verified at fixed intervals. The deviation between the compensated data and the data of the standard metering equipment is compared, and the error benchmark value is dynamically corrected.

[0016] Preferably, the existing data transmission channel includes at least one of RS485, power line carrier, low-power wireless and NB-IoT, and the adaptation method is selected according to the signal coverage capability of the installation scenario. For complex scenarios, a multi-channel redundant transmission design is adopted. The edge computing module is integrated into the existing terminal equipment of the metering box, without the need for additional hardware, and realizes local data processing and analysis through embedded algorithms; The local storage of 3D datasets adopts partition management, and is archived according to data type, acquisition time and importance. Data indexes are established to improve the efficiency of retrieval, comparison and traceability.

[0017] Preferably, the ventilation structure includes a door ventilation grille and a built-in ventilation fan. In the adaptive control logic, the ventilation duration is positively correlated with the temperature and humidity difference, and is dynamically adjusted in combination with the changing trend of environmental parameters. When the temperature and humidity difference is decreasing, shorten the ventilation time; when it is increasing, extend the ventilation time. The prediction model iteration cycle is set according to the regional climate stability. The cycle is shortened in areas with large climate fluctuations to adapt to environmental changes, while the cycle is extended in areas with stable climates to reduce computational resource consumption.

[0018] Preferably, in the dual-engine recognition mode, rule judgment completes the initial screening of anomalies based on preset parameter thresholds, and compares the deviation of real-time parameters with preset safety thresholds and normal fluctuation ranges; Trend analysis performs secondary verification by comparing historical data from the same period, long-term operating curves of the equipment, and data from similar equipment, eliminating misjudgments caused by accidental fluctuations and environmental interference. The combination of these two methods effectively reduces the false alarm rate of a single identification mode. The early warning information includes the time of the anomaly, characteristic parameters, associated device number, installation location, and preliminary handling suggestions, providing clear guidance for maintenance personnel to handle the situation on-site.

[0019] Preferably, the data platform supports hierarchical management of permissions for both power suppliers and users, assigning data viewing, operation, and export permissions through role division to ensure data security and information privacy; The operation and maintenance analysis report includes statistics on high-frequency problem areas, analysis of equipment failure patterns, evaluation of operation and maintenance efficiency, and suggestions for strategy optimization, providing data support for regional operation and maintenance management; The full lifecycle traceability covers all stages of data collection, condensation control, anomaly early warning, inspection and handling, and error compensation, establishing a data chain linking each stage to achieve data backtracking, problem tracing, and responsibility identification at any node.

[0020] Compared with existing technologies, the beneficial effects of this invention are as follows: it includes real-time acquisition and preprocessing of electrical, environmental, and status parameters to construct a three-dimensional dataset; calculation of condensation probability based on environmental parameters to trigger adaptive control; extraction of abnormal features for dual-engine mode identification and graded early warning; a status scoring matching inspection mechanism combined with the dataset; real-time compensation of measurement errors using preset temperature and humidity range benchmark values; and the establishment of a shared data platform to achieve full-process traceability. Through a dynamic adaptation mechanism, it solves the problems of inaccurate condensation prevention and control, high false alarm rate of abnormal identification, blind operation and maintenance scheduling, insufficient measurement accuracy, and poor data collaboration. It has the advantages of improving the accuracy of condensation prevention and control, reducing the false alarm rate of abnormal identification, optimizing the allocation of operation and maintenance scheduling resources, improving measurement accuracy, and enhancing data collaboration and traceability capabilities. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the operation and maintenance and metering optimization method according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 This paper presents a dynamic adaptive operation and maintenance (O&M) and metering optimization method for electricity metering boxes, aiming to address the aforementioned issues. The method collects multi-dimensional parameters in real time, performs local preprocessing and analysis via edge computing, and constructs a three-dimensional dataset encompassing electrical, environmental, and state aspects. Based on this dataset, the method can dynamically predict condensation risks and adaptively adjust existing ventilation structures, while iteratively optimizing model parameters. Furthermore, the method employs a dual-engine model to identify anomalies and provide tiered early warnings, combining comprehensive state scoring with inspection cycles and priorities to form a closed-loop O&M mechanism. The method also presets a metering error benchmark value, calculates compensation values ​​in real time, and remotely corrects electricity meter parameters. Finally, by establishing a shared data platform for both power supply and consumption, the method achieves full-process data traceability and automatic generation of O&M analysis reports, providing data support for strategy iteration.

[0024] For ease of understanding, the key terms in this embodiment are explained below: Sensing devices and data transmission channels refer to various sensors inside and outside the power metering box used to sense and acquire electrical, environmental and box status information, as well as communication links that transmit the sensed data to the processing unit.

[0025] Electrical parameters refer to indicators reflecting the internal electrical operating status of the electricity metering box, such as voltage, current, power factor, and energy data. Environmental parameters refer to the physical quantities of the internal and external environment of the electricity metering box, such as internal and external temperature and humidity, and atmospheric humidity. Box status parameters refer to indicators of the physical structure and operating status of the electricity metering box, such as the status of the door magnetic switch, vibration parameters, and humidity difference at the sealing points.

[0026] Noise reduction and normalization preprocessing refer to the process of cleaning and standardizing the raw collected data. Noise reduction aims to eliminate random errors and interference in the data; normalization aims to transform data with different dimensions to a uniform scale range.

[0027] The electrical-environmental-state 3D dataset refers to a multi-dimensional data set that integrates pre-processed electrical parameters, environmental parameters, and enclosure status parameters. This dataset comprehensively reflects the operational status of the electricity metering box, the relationship between its environment and physical state.

[0028] An edge computing module is a computing unit integrated into an electricity metering box or its nearby terminal equipment. This module has the ability to store data locally, perform preliminary analysis and processing, and reduce the bandwidth requirements and latency of remote data transmission.

[0029] A pre-defined predictive model refers to a mathematical model constructed based on historical data and specific algorithms to predict the probability of an event occurring. In this embodiment, the model is used to calculate the probability of condensation occurring.

[0030] Adaptive control logic refers to a control strategy that dynamically adjusts the operating status or parameters of equipment based on real-time monitoring data and preset rules.

[0031] The dual-engine mode of rule-based judgment and trend analysis refers to a mode that combines two data analysis methods for anomaly identification. The rule-based judgment engine performs initial screening based on preset thresholds or conditions; the trend analysis engine identifies patterns in data changes by comparing historical data or long-term operating curves, and performs secondary verification of anomalies.

[0032] A data sharing platform for both power suppliers and consumers refers to a platform that centrally manages and shares data related to electricity metering boxes. This platform enables data interoperability and collaboration, and supports data traceability, analysis, and report generation.

[0033] This embodiment provides a dynamically adaptable operation and maintenance and metering optimization method for electricity metering boxes, the specific implementation of which is as follows: First, relying on the existing sensing devices and data transmission channels of the electricity metering box, electrical parameters, environmental parameters, and box status parameters are collected in real time. This can be done through regular manual inspections, recording relevant parameters visually or with handheld devices. Alternatively, only electricity metering data can be collected, ignoring environmental or box status information. After acquiring the raw data, noise reduction and normalization preprocessing are performed. The unprocessed raw data can then be used directly for subsequent analysis, or only simple outlier removal can be performed. Subsequently, a three-dimensional dataset of electrical-environmental-status data is constructed, and the data is locally stored and preliminarily analyzed via an edge computing module, reducing remote transmission redundancy. Alternatively, all collected raw data can be directly uploaded to a remote server for centralized storage and processing, without local preliminary analysis.

[0034] Secondly, based on the pre-processed environmental parameters, the probability of condensation is calculated using a preset predictive model. Condensation risk can be assessed solely based on whether the internal temperature or humidity reaches a fixed threshold. Adaptive control logic is triggered by the difference in environmental parameters inside and outside the chamber, linking with the existing ventilation structure to achieve condensation prevention. The ventilation structure can be set to activate at fixed times each day, or only when the internal humidity exceeds a fixed value. Simultaneously, control data is recorded, and this data is used to iteratively optimize the model parameters. Alternatively, control actions can be recorded without feeding the control effects back to the model for optimization.

[0035] Furthermore, abnormal features are extracted from electrical parameters and enclosure status parameters to construct multi-dimensional feature vectors. This allows for monitoring only whether the instantaneous values ​​of voltage or current exceed preset ranges, without considering other parameters or their combinations. A dual-engine approach of rule-based judgment and trend analysis is employed to identify anomalies. It can rely solely on preset fixed rules for anomaly detection without comparing historical trends. Warning levels are categorized by severity, generating warning messages that are pushed to the maintenance terminal and linked to historical data comparison to improve accuracy. A unified warning message can be generated for all identified anomalies, regardless of severity, and only displayed via local indicator lights, without remote push notifications or historical data comparisons.

[0036] Furthermore, a comprehensive status score is performed on the metering boxes using a 3D dataset. A simple score can be given based solely on the metering box's years of operation or the date of its most recent maintenance. Equipment status levels are categorized according to the score, and corresponding inspection cycles and priorities are matched, forming a closed-loop operation and maintenance mechanism of scoring-scheduling-inspection-reassessment. A uniform periodic inspection cycle and content can be applied to all metering boxes, without differentiation based on their status levels, and without post-inspection reassessment.

[0037] In addition, preset metering error benchmark values ​​are provided for different temperature and humidity ranges. Alternatively, a single universal metering error benchmark value applicable to all temperature and humidity conditions can be preset. The edge computing module compares the internal temperature and humidity with the benchmark range in real time, calculates the error compensation value, and corrects the metering parameters via remote commands. Manual reading of the internal temperature and humidity is performed periodically, and the metering parameters are manually adjusted based on experience. Compensation accuracy is periodically verified to ensure metering accuracy. Accuracy verification is performed within the meter's calibration cycle, without periodic verification after compensation.

[0038] Finally, a shared data platform is established for both power suppliers and consumers to synchronize early warning information, inspection records, error compensation data, and control logs. Power suppliers and consumers can manage their data independently without information sharing. This enables full-process data traceability and automatically generates operation and maintenance analysis reports, providing data support for strategy iteration. However, only data from specific stages can be traced, and the operation and maintenance analysis reports are compiled manually on a regular basis and are not used to guide continuous strategy optimization.

[0039] This embodiment integrates multi-source data acquisition, edge computing, dynamic condensation prediction and adaptive control, multi-dimensional anomaly identification and hierarchical early warning, intelligent operation and maintenance scheduling, and real-time metering error compensation. It also constructs a shared data platform for both power supply and consumption parties. This solves the problems of traditional power metering boxes, such as passive and inefficient condensation control, high false alarm rates in anomaly identification, blind and wasteful operation and maintenance scheduling, lack of dynamic compensation for metering accuracy, and poor data collaboration and traceability. Therefore, it improves the operational stability and metering accuracy of power metering boxes, protects the legitimate rights and interests of both power supply and consumption parties, and enhances the safe and efficient operation of the power system.

[0040] In some of the solutions described above in this application, electrical parameters, environmental parameters, and enclosure status parameters are collected and preprocessed to construct a three-dimensional dataset. In this process, unclear parameter definitions may lead to incomplete data collection and failure to cover key influencing factors; fixed data collection modes may not be able to efficiently adapt to the changing characteristics of different parameters, resulting in wasted resources or delayed event response; insufficient preprocessing may introduce noise, missing values, or outliers, affecting the integrity and reliability of the dataset, thereby reducing the accuracy of subsequent analysis.

[0041] In response, this application further proposes a dynamic adaptive operation and maintenance and metering optimization method for the aforementioned electricity metering box. The electrical parameters include voltage, current, power factor, and energy data; environmental parameters include internal and external temperature and humidity, and atmospheric humidity; and box status parameters include door magnetic switch status, vibration parameters, and humidity difference at the seal. Electrical parameters are core indicators reflecting the internal electrical operation of the electricity metering box, such as voltage, current, power factor, and energy data. These parameters are directly related to electrical safety, power quality, and metering accuracy. Environmental parameters describe the physical environment inside and outside the electricity metering box, such as internal and external temperature and humidity, and atmospheric humidity. These parameters are important factors affecting equipment performance, condensation formation, and metering errors. Box status parameters focus on the physical integrity and safety of the electricity metering box, such as door magnetic switch status, vibration parameters, and humidity difference at the seal. These parameters help determine whether the box has been illegally opened, whether there is mechanical stress, or whether the seal has failed. By clearly defining these parameters, the comprehensiveness of data collection is ensured, laying the foundation for subsequent comprehensive analysis.

[0042] Data acquisition employs a combined timed and triggered approach. Electrical and environmental parameters are acquired on a timed basis, while door magnetic switch status and humidity difference at the seal are acquired on a triggered basis. Vibration parameters are acquired using a combined timed and triggered approach. The timed acquisition mode is suitable for parameters with relatively stable changes or requiring periodic monitoring, such as electrical and environmental parameters. Data can be read at preset time intervals, such as every minute or hour, ensuring data continuity and the effectiveness of trend analysis. The triggered acquisition mode is suitable for sudden, event-driven parameters, such as door magnetic switch status and humidity difference at the seal. Data is acquired only when a specific event occurs, avoiding unnecessary resource consumption and enabling timely response to anomalies. The combined acquisition mode combines the advantages of timed and triggered approaches. For example, for vibration parameters, in addition to timed acquisition, more intensive acquisition can be triggered when the vibration amplitude or frequency exceeds a preset threshold to capture detailed information about abnormal vibrations. This composite acquisition mode optimizes data acquisition efficiency based on the characteristics of different parameters, ensuring timely acquisition of critical information.

[0043] The preprocessing process also includes sensor interference removal, data completion, and outlier filtering. After verification using preset data validity thresholds, the integrity and reliability of the 3D dataset are ensured. Specifically, sensor interference removal involves identifying and eliminating erroneous data caused by sensor malfunctions, external electromagnetic interference, etc., using filtering algorithms or statistical methods. Data completion involves filling in missing data due to transmission interruptions, sensor malfunctions, etc., using interpolation, regression analysis, or prediction based on historical data to maintain dataset integrity. Outlier filtering involves identifying and removing data points that significantly deviate from the normal range. These outliers may be caused by transient failures, measurement errors, or extreme events, and are detected using statistical or machine learning methods. After the above processing, the processed data is verified again using preset data validity thresholds, such as setting reasonable ranges for parameters and limits on rates of change, to ensure that all data meet the expected quality standards, thereby guaranteeing the integrity and reliability of the 3D dataset.

[0044] Through the above technical solutions, this application effectively solves the problems of incomplete data, low acquisition efficiency, and poor dataset reliability by specifying parameter definitions, optimizing data acquisition strategies, and strengthening preprocessing procedures. First, by clarifying the specific content of electrical parameters, environmental parameters, and enclosure status parameters, it ensures that the acquisition scope covers key dimensions of electrical performance, environmental changes, and physical state, avoiding data omissions due to ambiguous definitions and thus supporting subsequent multidimensional analysis. Second, a combined timed and triggered data acquisition mode is adopted. Electrical and environmental parameters are acquired on a timed basis, adapting to the relatively stable changes in these parameters and reducing unnecessary frequent acquisition. Door magnetic switch status and humidity difference at the seal are acquired on a triggered basis, responding to sudden event-driven changes and avoiding response delays in fixed modes. Vibration parameters are acquired using a timed and triggered linkage method, balancing the needs of continuous monitoring and event triggering, and balancing resource consumption and data integrity. Finally, in the preprocessing process, sensor interference is removed to eliminate noise, missing data is supplemented to ensure continuity, outliers are filtered to eliminate erroneous data, and quality verification is performed using a preset data validity threshold. This ensures the integrity and reliability of the three-dimensional dataset, providing support for building a high-quality analytical foundation.

[0045] In some of the solutions mentioned above in this application, a preset prediction model and adaptive control logic are proposed to achieve condensation prevention and control. In this process, the model construction may lack scientific basis and dynamic adaptability, and cannot accurately adapt to the climate characteristics and installation scenarios of different regions, resulting in inaccurate condensation prediction. At the same time, the control logic may be too rigid and unable to dynamically adjust the ventilation strategy according to environmental changes, which may easily lead to energy waste or untimely prevention and control, affecting the overall effect and efficiency of condensation prevention and control.

[0046] In response, this application further proposes specific implementation methods for condensation control, including: a pre-set prediction model based on the dew point temperature calculation principle, combined with the temperature inside the chamber and the humidity difference between inside and outside the chamber to construct a multi-parameter correlation model, combined with historical condensation data, corresponding environmental parameters and control effects for iterative training, adapting to different regional climate characteristics and installation scenarios, and an adaptive control logic with the probability of condensation occurrence and the difference between parameters inside and outside the chamber as the core trigger conditions. When the probability reaches a preset threshold and the temperature and humidity difference meets the control requirements, the ventilation structure is activated and the running time is dynamically adjusted according to the size of the difference. When there is no obvious difference between parameters inside and outside the chamber but the humidity inside the chamber exceeds the standard, an intermittent ventilation mode is adopted, and the ventilation structure is cyclically started and stopped at preset intervals to balance the condensation control effect and energy consumption.

[0047] The pre-defined prediction model is based on the dew point temperature calculation principle. The dew point temperature is the critical temperature at which water vapor in the air reaches saturation and begins to condense into liquid water. Based on this principle, the conditions for condensation to occur can be accurately predicted from a physics perspective. This principle can be implemented by measuring the temperature and relative humidity of the air inside the chamber and using the dew point temperature formula to calculate the dew point temperature in real time under the current environment. Alternatively, a lookup table can be created to directly look up the corresponding dew point temperature based on the chamber temperature and relative humidity, simplifying the real-time calculation process. Furthermore, the model combines the chamber temperature and the humidity difference between inside and outside the chamber to construct a multi-parameter correlation model, comprehensively considering multiple key environmental parameters to more comprehensively and accurately assess the risk of condensation. The chamber temperature directly affects the saturated water vapor content of the air, while the humidity difference between inside and outside the chamber reflects the driving force for water vapor to penetrate into the chamber or diffuse outwards, and is an important factor in condensation formation. The multi-parameter correlation model can be constructed using a multiple linear regression model, with the chamber temperature and humidity difference as input variables and the probability of condensation as the output, and the model parameters obtained through training with historical data. Alternatively, a neural network model can be used to construct a multilayer perceptron. Input parameters such as the temperature inside the chamber and the humidity difference between inside and outside the chamber are input, and the model is trained using deep learning algorithms to capture more complex nonlinear relationships. Decision tree or random forest models can also be used to divide condensation risk areas based on different parameter combinations. To ensure the model's dynamic adaptability, iterative training is conducted using historical condensation data, corresponding environmental parameters, and control effects to adapt to different regional climate characteristics and installation scenarios. Model training and optimization is a continuous process. By continuously learning from historical data and control feedback, the model can adapt to various complex and changing environments. Historical condensation data provides real-world examples of condensation occurrences, environmental parameters record the environmental conditions at the time, and control effects evaluate the effectiveness of previous control measures. This can be achieved by using online learning or incremental learning algorithms to periodically input new condensation event data, environmental data, and control results into the model for retraining, updating model weights or parameters. Alternatively, a regional climate database can be established, and independent model parameter sets can be trained and maintained for different geographical regions and installation scenarios, or transfer learning techniques can be used for model adaptation.

[0048] Meanwhile, the adaptive control logic of this application uses the probability of condensation and the difference in parameters inside and outside the chamber as the core triggering conditions. By comprehensively judging the probability of condensation and the degree of environmental difference, it ensures the precise activation of control measures. The probability of condensation directly quantifies the risk, while the difference in parameters inside and outside the chamber provides the strength of the physical driving force for condensation formation. A condensation probability threshold can be set, for example, above 70%, and a temperature and humidity difference threshold between inside and outside the chamber can be set, where the humidity inside the chamber is 5% RH higher than the humidity outside and the temperature inside the chamber is lower than the dew point temperature. Control is triggered when both conditions are met simultaneously. Alternatively, a fuzzy logic controller can be used, taking the probability of condensation and the difference in parameters inside and outside the chamber as fuzzy inputs, and using fuzzy rules to deduce the control intensity or whether to activate control. When the probability reaches the preset threshold and the temperature and humidity difference meets the control requirements, the ventilation structure is activated and the running time is dynamically adjusted according to the difference. This is the specific execution step for condensation control, ensuring timely activation of ventilation when necessary, and optimizing the effect and energy consumption by dynamically adjusting the ventilation duration. When the probability of condensation exceeds a preset threshold (e.g., 80%) and the humidity difference between the inside and outside of the chamber is such that the humidity inside the chamber is 10% RH higher than the humidity outside, meeting the control requirements, the ventilation system is immediately activated. The ventilation duration can be extended proportionally according to the magnitude of the temperature and humidity difference; the larger the difference, the longer the ventilation time. Alternatively, the recommended ventilation duration can be calculated using a lookup table or preset function based on the probability of condensation and the temperature and humidity difference, and then an instruction can be sent to the ventilation system for execution. Furthermore, for special operating conditions where there is no significant difference in parameters between the inside and outside of the chamber but the humidity inside the chamber exceeds the standard, this application adopts an intermittent ventilation mode, cyclically starting and stopping the ventilation system at preset intervals to balance the condensation control effect with energy consumption. This mode aims to gradually reduce the humidity inside the chamber and avoid energy waste caused by continuous ventilation when the difference between the external environment and the inside of the chamber is not significant, but the humidity inside the chamber is still too high. When the humidity difference between the inside and outside of the chamber is below a certain threshold (e.g., less than 3% RH) but the humidity inside the chamber is above a safe threshold (e.g., above 85% RH), intermittent ventilation is activated, with ventilation for 5 minutes every 30 minutes, cyclically executed until the humidity inside the chamber drops to a safe range. Alternatively, the cycle and duration of intermittent ventilation can be dynamically adjusted based on the degree of humidity exceeding the standard inside the chamber.

[0049] Through the above technical solution, this application effectively solves the problems of existing condensation prediction models lacking scientific basis and dynamic adaptability, and rigid control logic leading to energy waste or untimely prevention and control. Based on the dew point temperature calculation principle and multi-parameter correlation model, combined with real-time collected environmental parameters such as the temperature inside the chamber and the humidity difference between inside and outside the chamber, the risk of condensation can be predicted scientifically and accurately. At the same time, through iterative training of historical condensation data and control effects, the prediction model can continuously learn and optimize, dynamically adapting to the climate characteristics and installation scenarios of different regions, improving the accuracy of condensation prediction and the adaptability of the model. On this basis, the adaptive control logic uses the probability of condensation and the difference between parameters inside and outside the chamber as the core triggering conditions, ensuring that the ventilation structure is activated only when necessary, avoiding unnecessary energy consumption. By dynamically adjusting the ventilation operation time according to the difference, fine-grained control is achieved, which can effectively prevent and control condensation while maximizing energy conservation. Especially in special cases where there is no significant difference between parameters inside and outside the chamber but the humidity inside the chamber exceeds the standard, an intermittent ventilation mode is adopted to further optimize energy utilization efficiency while ensuring the effectiveness of condensation prevention and control. The synergistic effect of these technologies makes condensation control in electricity metering boxes more intelligent, efficient, and energy-saving, effectively protecting the electrical components inside the box, extending equipment life, and ensuring the accuracy of electricity metering and the safe and stable operation of the power system. Simultaneously, combined with the local processing capabilities of the edge computing module, rapid response to condensation prediction and control can be achieved, further improving the real-time performance and reliability of the entire operation and maintenance system.

[0050] In some of the embodiments described above in this application, abnormal features are proposed to identify anomalies and classify warning levels. However, in the implementation process, the specific definitions of abnormal features are not clear enough, and the classification of warning levels is not precise enough. This may lead to insufficient accuracy in anomaly identification and processing efficiency, and an inability to effectively distinguish between serious problems and potential hazards, thereby affecting the timeliness of warning information and the allocation of processing priorities.

[0051] In response, this application further proposes that the abnormal characteristics include the parameter mutation rate, voltage distortion rate, current fluctuation range, and power factor deviation of electrical parameters, as well as the abnormal opening time of the door magnet, vibration mutation amplitude, vibration duration, and humidity difference mutation value at the sealing point of the enclosure status parameters; the warning levels are at least divided into Level 1 and Level 2. Level 1 anomalies correspond to serious problems affecting electricity safety and metering accuracy, such as suspected electricity theft, line overload, and short circuit hazards. Level 2 anomalies correspond to potential hazards such as poor line contact, aging enclosure seals, and slight vibration interference; different levels of warning information correspond to different push priorities and processing procedures. Level 1 anomaly information is pushed to the operation and maintenance terminal in real time and triggers audible and visual reminders. Level 2 anomaly information is included in the operation and maintenance to-do list and pushed according to priority.

[0052] The definition of these abnormal features aims to comprehensively capture various abnormal situations during the operation of the electricity metering box. Specifically, the parameter mutation rate of electrical parameters refers to the magnitude of change in electrical parameters such as voltage and current over a short period. This can be achieved by differential calculation of continuously collected electrical parameters or by judging the deviation between the average value of a sliding window and the current value. Voltage distortion rate reflects the non-sinusoidal degree of the grid voltage waveform. It can be calculated by analyzing the harmonic content in the voltage waveform using Fourier Transform (FFT) or by real-time monitoring using a dedicated voltage distortion measurement module. Current fluctuation range refers to the difference between the maximum and minimum values ​​of the current within a certain time window. This can be obtained by statistically analyzing the extreme values ​​of current data or calculating the standard deviation. Power factor deviation value indicates the degree of deviation between the actual power factor and the ideal value, such as 0.9 or 1.0. This can be determined by measuring the power factor in real time and comparing it with a preset benchmark value. The door magnetic abnormal opening duration of the box status parameters refers to the duration during which the door magnetic switch is in an abnormally open state. This can be recorded by combining the door magnetic sensor status signal with a timer, or by judging unauthorized opening and timing it using a state machine model. Vibration abrupt change amplitude refers to a sudden increase or decrease in the vibration intensity of the enclosure. This can be identified by real-time vibration data acquisition using an accelerometer, comparing it with historical data or a preset threshold, and employing peak detection or root mean square (RMS) rate of change detection. Vibration duration refers to the length of time a vibration event lasts from start to finish. This can be achieved by triggering a timer with a vibration sensor and recording the duration, or by recording the start and end timestamps of the vibration event in an event log. Humidity difference abrupt change at the seal refers to a drastic change in the humidity difference between the inside and outside of the enclosure's sealing area. This can be identified by placing humidity sensors inside and outside the seal, calculating the humidity difference in real-time, and monitoring whether its rate of change exceeds a preset threshold, or by analyzing the short-term fluctuation trend of the humidity difference.

[0053] The warning levels are divided into at least Level 1 and Level 2, aiming to classify and manage anomalies according to their severity and potential impact. Level 1 anomalies typically address issues that have a direct and serious impact on electricity safety and metering accuracy, such as suspected electricity theft, manifested as abnormal electricity usage patterns or sudden changes in power factor; line overload, manifested as prolonged current exceeding limits or abnormal temperature increases; and short circuit hazards, manifested as sudden voltage drops or surges in current. Level 2 anomalies correspond to potential, non-immediate risks, such as poor line contact, manifested as localized temperature rises or slight voltage fluctuations; aging enclosure seals, manifested as persistently high humidity inside the enclosure or abnormal humidity differences at the seals; and minor vibration interference, manifested as vibration parameters slightly exceeding the normal range. The warning levels can be implemented based on an expert experience rule base, mapping the above-mentioned anomaly characteristics to the corresponding warning levels, or through machine learning models trained using historical anomaly data to automatically classify identified anomalies.

[0054] Different levels of early warning information correspond to different push priorities and processing procedures to ensure that resources are allocated rationally and utilized efficiently. Level 1 anomaly information, due to its severity, is pushed to the maintenance terminal in real time via SMS and mobile application notifications, triggering audible and visual alerts, such as warnings via local buzzers or indicator lights on the metering box, and simultaneously highlighted on the monitoring screen in the maintenance center. Level 2 anomaly information is included in the maintenance to-do list and pushed according to priority, via email notifications, work order system task assignments, or displayed with lower priority in the maintenance management platform, allowing maintenance personnel to handle it according to their work schedule. This differentiated push mechanism ensures the fastest possible response to emergencies, while effectively managing potential problems.

[0055] Through the above technical solution, this application effectively solves the problems of vague anomaly feature definitions and imprecise warning level classifications. It specifically defines the anomaly features of electrical parameters and enclosure status parameters, making anomaly identification more comprehensive and capable of capturing subtle anomalies that are difficult to detect using traditional methods, thereby improving the accuracy of anomaly identification. Simultaneously, it clearly divides warning levels into Level 1 and Level 2, and details the specific problems corresponding to each level, enabling maintenance personnel to clearly judge the severity and potential impact of anomalies. This refined hierarchical management, combined with differentiated push priorities and processing procedures corresponding to different levels of warning information, ensures that serious problems receive timely and prioritized responses and handling, while potential hazards can be incorporated into daily maintenance plans, avoiding resource waste and processing delays. For Level 1 anomalies such as suspected electricity theft or short circuit hazards, the system can immediately trigger audible and visual alerts and push them to the maintenance terminal in real time, prompting maintenance personnel to intervene immediately and effectively preventing the expansion of electrical safety accidents and metering losses. For level-two anomalies such as aging of the enclosure seal, they are included in the to-do list, allowing maintenance personnel to investigate them during routine inspections or planned maintenance, thereby optimizing the allocation of maintenance resources and improving the overall maintenance efficiency and accuracy.

[0056] In some of the embodiments described above in this application, a comprehensive status score is proposed to classify equipment status levels and match inspection cycles. However, in its implementation, the scoring method lacks quantitative standards, the weight allocation is fixed and does not highlight key factors, and it cannot be dynamically adjusted according to actual application scenarios, resulting in inaccurate status assessment, unreasonable matching of inspection cycles and priorities, and waste of resources or omission of hidden dangers.

[0057] In response, this application further proposes that the comprehensive status score adopts a weighted quantitative evaluation, assigning corresponding weights to the stability of measurement accuracy, environmental adaptability, and physical status of the enclosure, with the stability of measurement accuracy having the highest weight, while the weights of the other two can be dynamically adjusted according to the application scenario; the equipment is divided into three levels: excellent, qualified, and warning, based on the score. The inspection cycle is extended for excellent-level equipment, the inspection cycle is maintained at the regular level with a focus on key warning items, and the inspection cycle is shortened for warning-level equipment with priority given to on-site maintenance; after the inspection is completed, the status score is updated based on the on-site testing data, forming a closed-loop evaluation system.

[0058] Specifically, comprehensive status scoring is a method for quantitatively evaluating the overall operational status of an electricity metering box. Using a weighted quantitative assessment means assigning different importance coefficients (weights) to different assessment dimensions, and then summing the scores of each dimension using weighted averages to obtain an overall quantitative score. Its purpose is to provide an objective and comparable indicator of equipment status, avoiding subjective judgment and improving the accuracy and scientific rigor of the assessment. Initial weights can be determined using expert scoring combined with the Analytic Hierarchy Process (AHP), followed by iterative optimization through historical data backtesting and machine learning algorithms. Alternatively, a fuzzy comprehensive evaluation method can be used, performing fuzzy operations on the fuzzy membership degrees of each assessment indicator and the weight vector to obtain the comprehensive evaluation result.

[0059] This application assigns corresponding weights to metering accuracy stability, environmental adaptability, and enclosure physical condition. Metering accuracy stability is a core functional indicator used to evaluate the metering accuracy and drift of the electricity metering box under different operating conditions. Environmental adaptability assesses the metering box's adaptability and protection against external environmental factors such as temperature, humidity, and condensation risks. Enclosure physical condition assesses the metering box's structural integrity, sealing, and anti-theft / vandalism capabilities. These weights can be preset as fixed proportions: metering accuracy stability 50%, environmental adaptability 30%, and enclosure physical condition 20%; alternatively, the weights can be manually adjusted via the user interface or configuration tools, or dynamically adjusted based on historical fault data and maintenance experience.

[0060] In this application, the weight of metering accuracy and stability is the highest, while the weights of the other two dimensions can be dynamically adjusted according to the application scenario. This emphasizes the core position of metering accuracy in the operation and maintenance of electricity metering boxes, ensuring that the evaluation results prioritize the key indicator of metering accuracy. Simultaneously, the weights of other dimensions can be adjusted according to actual application scenarios; for example, in high-humidity areas, environmental adaptability is emphasized, while in areas susceptible to external damage, the physical condition of the box is emphasized, enhancing the flexibility and applicability of the evaluation model. Dynamic fine-tuning can be achieved through the backend management system, automatically loading preset weight configurations based on the user-selected region type (coastal, mountainous, urban) or equipment type (outdoor, indoor). Alternatively, machine learning models can automatically learn and adjust weights based on historical operation and maintenance data, such as fault types and environmental conditions, to maximize predictive accuracy.

[0061] Equipment is categorized into three levels—Excellent, Satisfactory, and Warning—based on its scoring. This aims to discretize the continuous overall status score into easily understandable and manageable levels, providing a clear decision-making basis for subsequent operation and maintenance scheduling. Fixed scoring thresholds can be set, such as 90 points or above for Excellent, 70-90 points for Satisfactory, and below 70 points for Warning. Alternatively, the thresholds can be dynamically determined based on historical data analysis using statistical methods such as quantiles to adapt to the overall performance of different batches of equipment.

[0062] Based on equipment status levels, this application proposes extending the inspection cycle for excellent-level equipment, maintaining the regular cycle for qualified-level equipment while focusing on key early warning items, and shortening the inspection cycle for early warning-level equipment while prioritizing on-site maintenance. This is a differentiated maintenance strategy aimed at optimizing inspection resource allocation and improving maintenance efficiency and accuracy. For excellent-level equipment, resources can be optimized to reduce unnecessary inspections. For qualified-level equipment, basic inspections are maintained, but with greater attention paid to potential risks. For early warning-level equipment, a rapid response is implemented, prioritizing high-risk equipment. Inspection cycles and key inspection items can be preset in the maintenance management system and automatically linked to equipment levels. For example, if the inspection cycle for excellent-level equipment is extended from monthly to quarterly, early warning equipment will immediately trigger emergency inspection tasks. Alternatively, intelligent scheduling algorithms can be used to automatically generate optimal inspection routes and task allocations based on the geographical location, skills, and urgency of the equipment used by maintenance personnel.

[0063] After inspection, the status score is updated based on on-site testing data, forming a closed-loop evaluation system. This emphasizes the dynamism and adaptability of the evaluation system. On-site testing data, such as maintenance records, component replacements, and actual measurements, directly reflect the true state of the equipment. Incorporating this data into the scoring update mechanism can correct biases in the evaluation model, making the scoring results more realistic and providing data support for subsequent strategy optimization. The closed-loop system ensures continuous improvement across evaluation, scheduling, execution, and feedback. After completing on-site inspections or maintenance, maintenance personnel input testing data and processing results via mobile terminals. The system automatically triggers the recalculation and update of the status score. Alternatively, it can integrate data interfaces with on-site testing equipment, such as infrared thermometers and insulation testers, to achieve automatic data uploading and score updates.

[0064] Through the above technical solution, this application solves the problems of traditional operation and maintenance (O&M) scoring methods lacking quantitative standards and having fixed weight allocations that cannot be dynamically adjusted. The weighted quantitative assessment makes the comprehensive status score more objective and scientific, avoiding biases from subjective judgment. Metering accuracy and stability have the highest weight, ensuring priority focus on the core functions of the electricity metering box, while the other two weights can be dynamically fine-tuned according to application scenarios, enhancing the adaptability of the assessment model to different regional climates, installation environments, and other actual conditions, thus improving the accuracy of the assessment. Dividing equipment into three levels—excellent, qualified, and warning—achieves refined and differentiated O&M scheduling. Excellent-level equipment has extended inspection cycles, effectively reducing excessive inspections of stable equipment and saving human and material resources. Warning-level equipment has shortened inspection cycles and is prioritized for on-site O&M, ensuring timely response and handling of high-risk equipment and preventing the expansion of potential hazards. After inspection, the status score is updated based on on-site testing data, forming a closed-loop O&M mechanism of assessment-scheduling-inspection-reassessment. This allows for continuous correction and optimization of the status assessment results, improving the adaptability and effectiveness of the O&M strategy. By combining the real-time collected electrical parameters, environmental parameters, and box status parameters in the above-mentioned dynamic adaptive operation and maintenance and metering optimization method for electricity metering boxes, as well as the preliminary analysis by the edge computing module, the data source for the comprehensive status score becomes more comprehensive, real-time, and reliable, further improving the accuracy of the assessment and the level of intelligent operation and maintenance.

[0065] In some of the solutions mentioned above in this application, a measurement error benchmark value is proposed to compensate for measurement errors caused by changes in temperature and humidity in real time. However, due to the significant differences in error characteristics of different meter box types, such as single-phase meters and three-phase meters, the preset benchmark value may not be differentiated for different types, resulting in insufficient compensation accuracy. The acquisition of the benchmark value may only rely on laboratory testing without combining actual on-site operating data, affecting applicability. Furthermore, the lack of a dynamic verification mechanism after compensation may lead to error accumulation and affect long-term accuracy.

[0066] In response, this application further proposes to pre-set the metering error benchmark value according to the meter box type, covering common types such as single-phase meters, three-phase meters, and meter boxes connected via current transformers. Each type corresponds to an error benchmark value in different temperature and humidity ranges. The benchmark value is obtained by combining laboratory standard environment simulation testing with calibration based on actual on-site operating data to ensure accuracy and applicability. Error compensation does not change the hardware structure of the energy meter, but is only achieved by adjusting the built-in software metering coefficient through remote commands. After compensation, the accuracy is verified at fixed intervals, and the deviation between the compensated data and the standard metering equipment data is compared to dynamically correct the error benchmark value.

[0067] The metering error benchmark values ​​are preset according to the metering box type, covering common types such as single-phase meters, three-phase meters, and metering boxes connected via current transformers. Each type corresponds to an error benchmark value for different temperature and humidity ranges, aiming to provide more targeted error compensation basis for different types of energy metering boxes. A multi-dimensional lookup table can be established in the system database, with dimensions including metering box type (e.g., single-phase, three-phase, current transformer connected) and temperature and humidity range (e.g., -20℃~0℃, 0℃~20℃, 20℃~40℃), with each intersection storing the corresponding error benchmark value. When the edge computing module obtains the current metering box type and temperature and humidity data, it can quickly locate and obtain the corresponding error benchmark value through this lookup table. A set of parameterized error functions can be preset for different metering box types. This function takes temperature and humidity as input variables and outputs the corresponding error benchmark value. In practical applications, the corresponding error function is selected according to the metering box type, and the accurate error benchmark value is calculated by substituting real-time temperature and humidity data.

[0068] The benchmark values ​​are obtained through a combination of laboratory standard environment simulation testing and calibration using actual field operation data, ensuring accuracy and applicability. This ensures that the benchmark values ​​possess both high accuracy under controlled environments and adaptability to the complexity of actual operating environments. Specifically, firstly, in a laboratory with temperature and humidity control capabilities, multi-point temperature and humidity conditions are used to test the metering performance of different types of electricity meters, recording their error data under standard conditions to form initial benchmark values. Subsequently, these electricity meters are deployed in actual operating environments, and their error data under actual operating conditions is collected through synchronous comparison with high-precision standard meters. Statistical analysis methods are then used to correct and optimize the initial benchmark values ​​to reflect the complexity of the field environment. Another approach is to establish a continuously learning benchmark calibration system, initially obtaining preliminary benchmark values ​​through laboratory testing. After the electricity metering boxes are put into operation, the system periodically or when triggered by specific events collects actual error data from the field metering boxes and transmits this data back to the central server. The server uses machine learning algorithms, combined with laboratory and field data, to iteratively train and calibrate the benchmark model, enabling it to better adapt to the actual operating conditions of different regions and seasons.

[0069] Error compensation does not alter the hardware structure of the electricity meter; it is achieved solely through remote commands adjusting the built-in software metering coefficients. After compensation, accuracy is verified periodically, comparing the compensated data with that of a standard metering device to dynamically correct the error benchmark value. This compensation method is non-intrusive, flexible, and allows for continuous optimization. The edge computing module generates specific remote commands based on the calculated error compensation value and sends them to the electricity meter via a data transmission channel. Upon receiving the commands, the meter's built-in firmware or software modifies its internally stored metering coefficients, including current transformer ratios, voltage transformer ratios, or energy accumulation coefficients, automatically applying the compensation in subsequent electricity metering processes. This adjustment is software-level and does not involve changes to the physical circuitry. After the compensation operation is complete, the system initiates a verification cycle, monthly or quarterly. During this cycle, the edge computing module continuously monitors the electrical and environmental parameters of the metering box and periodically compares the data with a known, more accurate standard metering device. This could be a field-deployed reference meter or cross-referencing with other stably operating metering boxes in the same area via remote data analysis. The comparison results generate a deviation report to evaluate the compensation effect. If the verification results show that the compensated data still deviates from the standard metrology equipment data, or if the deviation shows a new trend, the system will trigger a dynamic correction process for the error benchmark value. This can be achieved by feeding back the verification deviation data to the benchmark value acquisition module, allowing it to retrain or adjust its parameters, thereby updating the error benchmark value for that phenotype or that specific metrology chamber within a specific temperature and humidity range.

[0070] Through the above technical solutions, the metering error benchmark value can be precisely preset for different metering box types and temperature and humidity ranges, effectively solving the problem of insufficient compensation accuracy of general benchmark values ​​under different equipment characteristics. Simultaneously, by combining the accuracy of laboratory testing with the practicality of on-site operating data to obtain the benchmark value, the accuracy and on-site applicability of the benchmark value are improved, ensuring the reliability of compensation. Furthermore, the non-intrusive compensation method of adjusting the built-in software metering coefficients via remote commands avoids the complexity and cost of hardware modifications, achieving efficient, flexible, and economical compensation. After compensation, the accuracy is verified at fixed intervals, and the error benchmark value is dynamically corrected, ensuring real-time monitoring of the compensation effect and deviation identification, preventing error accumulation, and making the compensation mechanism adaptive and long-term stable. Overall, these technical features work synergistically to solve the problems of inaccurate benchmark value settings, single acquisition methods, and lack of verification after compensation, improving the metering accuracy of the electricity metering box and the sustainability of the system.

[0071] In some of the embodiments described above in this application, data acquisition and preprocessing methods are proposed to construct three-dimensional datasets and ensure their integrity and reliability. However, in the implementation process, the selection of data transmission channels lacks flexibility and cannot adapt to the signal coverage capabilities of different installation scenarios, resulting in unstable or interrupted data transmission. Edge computing modules may require additional hardware, increasing system costs and compatibility risks. Local storage management is chaotic and disorderly, affecting the efficiency of data retrieval, comparison, and tracing, and hindering the rapid access and analysis of three-dimensional datasets.

[0072] In response, this application further proposes technical solutions for optimizing data transmission, edge computing module integration, and local storage management. Specifically, existing data transmission channels include at least one of RS485, power line carrier, low-power wireless, and NB-IoT. The adaptation method is selected according to the signal coverage capability of the installation scenario. For complex scenarios, a multi-channel redundant transmission design is adopted. The edge computing module is integrated into the existing terminal equipment of the metering box without the need for additional hardware. Local data processing and analysis are achieved through embedded algorithms. The local storage of the 3D dataset adopts partition management, classifying and archiving it according to data type, collection time, and importance, and establishing a data index to improve the efficiency of retrieval, comparison, and traceability.

[0073] The existing data transmission channels are designed to provide diverse communication options to adapt to different field environments. RS485, as an industrial-grade serial communication standard, features moderate transmission distance, strong anti-interference capabilities, and support for multi-point communication, making it suitable for wired connections within metering boxes or between short-distance devices. Power line carrier technology utilizes existing power lines for data transmission, eliminating the need for additional wiring, making it particularly suitable for areas with well-developed power infrastructure and effectively reducing deployment costs. Low-power wireless technologies, such as LoRa and Zigbee, offer advantages such as low power consumption, wide coverage, and low cost, making them suitable for scenarios requiring wireless transmission but with relatively small data volumes. NB-IoT, a cellular IoT technology, provides wide-area coverage and low-power connectivity, suitable for remote areas or scenarios requiring carrier network support. By selecting the adaptation method based on the signal coverage capabilities of the installation scenario, NB-IoT or low-power wireless can be prioritized in urban areas with good signal coverage, while RS485 or power line carrier can be selected in signal blind spots or areas with strong interference. In complex scenarios, a multi-channel redundant transmission design can be adopted, which can simultaneously enable NB-IoT as the main channel and use power line carrier as the backup channel, or automatically switch to the backup channel when the main channel fails, thereby ensuring the continuity and reliability of data transmission.

[0074] The core of the edge computing module lies in bringing data processing capabilities down to the data source, namely, inside the metering box. This module is integrated into the existing terminal equipment within the metering box, embedded as a software module into the existing processor of the smart meter, data collector, or communication module, or directly integrated onto the existing circuit board as a micro-computing unit. This avoids the need for additional hardware, effectively controlling system costs and reducing compatibility risks. Local data processing and analysis are achieved through embedded algorithms, which can be pre-defined rule engines, lightweight machine learning models, or statistical analysis models. These algorithms perform noise reduction, normalization preprocessing, feature extraction, condensation probability calculation, and preliminary anomaly detection on the collected raw data locally, thereby reducing reliance on remote servers, lowering data transmission volume, and ensuring timely data processing.

[0075] The local storage of the 3D dataset aims to optimize data organization and management. A partitioned management approach logically divides the storage space into different areas, such as electrical parameter data areas, environmental parameter data areas, enclosure status parameter data areas, control log areas, and early warning information areas. Data is categorized and archived according to data type, acquisition time, and importance. Electrical data such as voltage, current, and power factor are grouped into one category and archived by hour or day. Environmental data such as temperature and humidity, and atmospheric humidity are grouped into another category, also archived by time. Status data such as door magnetic switch status and vibration parameters are grouped into another category. For highly important data, such as real-time anomaly early warning data, higher storage priority and faster access paths can be set. Data indexing is established using database indexing technologies such as B-tree indexes and hash indexes. Indexes are created for key fields such as timestamps, device IDs, and data types, thereby improving the efficiency of data retrieval, comparison, and tracing, and providing fast and accurate data support for subsequent data analysis and decision-making.

[0076] Through the above technical solutions, this application effectively solves the problems of inflexible data transmission channel selection, high hardware costs for adding edge computing modules, and chaotic local storage management. Specifically, by providing multiple data transmission channels and adapting them to the signal coverage capabilities of the installation scenario, and employing a multi-channel redundant transmission design in complex scenarios, the flexibility, stability, and reliability of data transmission are greatly enhanced. This ensures continuous and complete data acquisition in various environments and avoids data interruption or loss due to signal problems. Integrating the edge computing module into the existing terminal equipment of the metering box eliminates the need for additional hardware, reducing system deployment costs and compatibility risks. Embedded algorithms enable real-time local data processing and preliminary analysis, effectively reducing remote transmission redundancy and improving the timeliness and efficiency of data processing. Furthermore, the local storage of the 3D dataset adopts partitioned management and is archived according to data type, acquisition time, and importance. A data index is also established, making local data management more orderly and efficient. This greatly improves the efficiency of data retrieval, comparison, and traceability, providing fast and reliable data support for subsequent condensation prevention, anomaly identification, status scoring, and metering optimization, thereby comprehensively improving the operation and maintenance and metering optimization capabilities of the electricity metering box.

[0077] In some of the solutions mentioned above in this application, ventilation structures are proposed to achieve condensation control. In this process, the ventilation duration may not be dynamically optimized and adjusted according to the changing trend of environmental parameters, resulting in increased energy consumption or unsatisfactory condensation control effect. When the changing trend of temperature and humidity difference is not taken into account, fixed ventilation mode is prone to waste of resources or untimely control.

[0078] In response, this application further proposes that the ventilation structure includes a door ventilation grille and a built-in ventilation fan. In the adaptive control logic, the ventilation duration is positively correlated with the temperature and humidity difference, and is dynamically adjusted in combination with the trend of environmental parameter changes. When the temperature and humidity difference shows a decreasing trend, the ventilation duration is shortened; when it shows an increasing trend, the ventilation duration is extended. The prediction model iteration cycle is set according to the regional climate stability. In areas with large climate fluctuations, the cycle is shortened to adapt to environmental changes, and in areas with stable climates, the cycle is extended to reduce computational resource consumption.

[0079] Specifically, the ventilation structure is a device used to promote air exchange between the inside and outside of the electricity metering box. Its function is to reduce humidity inside the box and dissipate heat, thereby effectively preventing condensation. This ventilation structure can include a door ventilation grille and a built-in ventilation fan. The door ventilation grille can be a louvered, mesh, or perforated plate design, typically installed on the door or side wall of the box, achieving air circulation through natural convection. Its opening area, shape, and position can be optimized according to the box size and expected ventilation volume. The built-in ventilation fan can be an axial fan, centrifugal fan, or crossflow fan, installed inside the box, accelerating air exchange through forced convection. The fan size, speed, and power can be selected according to the internal space, required airflow, and power consumption requirements. These two structures can be used individually or in combination; the ventilation grille provides basic passive ventilation, while the ventilation fan provides active, enhanced ventilation when needed.

[0080] Based on this, the adaptive control logic of this application shows a positive correlation between ventilation duration and the temperature and humidity difference, while also dynamically adjusting based on the changing trends of environmental parameters. This adaptive control logic refers to an intelligent control strategy that automatically adjusts the operating time of the ventilation structure based on real-time monitored environmental parameters, particularly the temperature and humidity difference and its changing trends. This aims to ensure the accuracy and efficiency of ventilation operations and avoid over- or under-ventilation. The logic can be set as follows: the larger the temperature and humidity difference between the inside and outside of the enclosure, the higher the risk of condensation, and the longer the cumulative operating time of the ventilation structure, such as the built-in ventilation fan. This can be achieved by mapping the difference to the ventilation duration through linear or nonlinear functional relationships. Furthermore, the logic can further incorporate the rate or direction of change of environmental parameters as adjustment factors. If the temperature and humidity difference continues to widen, even if the current difference has not reached a high level, ventilation can be brought forward or extended to address the impending high risk of condensation. This can be achieved by trend prediction based on historical data or by real-time monitoring of the slope of parameter changes.

[0081] Furthermore, when the temperature and humidity difference shows a decreasing trend, the system shortens the ventilation time; when the temperature and humidity difference shows an increasing trend, the system extends the ventilation time. This is a specific implementation detail of the adaptive control logic, designed to finely manage ventilation resources based on the dynamic changes in condensation risk, thereby optimizing energy consumption and improving the response speed and accuracy of condensation prevention and control. When the system detects that the temperature and humidity difference between the inside and outside of the chamber shows a decreasing trend over several consecutive sampling periods, the control logic will correspondingly reduce the duration of the current or next ventilation cycle; conversely, when the difference shows an increasing trend, the ventilation time will be increased. This trend judgment can be achieved by comparing the current difference with the average difference of the previous moment or several previous moments. A threshold for trend change and a corresponding duration adjustment gradient can also be set. When the decreasing trend of the difference exceeds a certain percentage, the ventilation time is reduced by a preset value; when the increasing trend of the difference exceeds a certain percentage, the ventilation time is increased by a preset value.

[0082] Furthermore, the prediction model iteration cycle is set according to regional climate stability. The prediction model iteration cycle refers to the time interval for updating model parameters or retraining the model used for condensation risk prediction. This cycle is dynamically adjusted based on the climate characteristics of the installation area to ensure the accuracy and adaptability of the prediction model, while optimizing the use of computing resources. Regional climate stability can be assessed by analyzing historical meteorological data, including the annual average temperature and humidity fluctuation range, the frequency of extreme weather events, and seasonal variation patterns. Areas with annual temperature and humidity variations below a certain threshold are defined as climate-stable areas, while those exceeding a certain threshold are considered areas with significant climate fluctuations.

[0083] Specifically, for regions with significant climate fluctuations, the system shortens the iteration cycle of the prediction model to adapt to environmental changes. This is a model iteration strategy for climate-unstable regions, aiming to improve the model's responsiveness and prediction accuracy in rapidly changing environments. In areas identified as having significant climate fluctuations, the system automatically increases the frequency of parameter updates or retraining of the condensation prediction model. This can be shortened from the usual quarterly iterations to monthly iterations, or immediately triggered for small-scale fine-tuning or local retraining when climate anomalies such as continuous rainfall or a sudden drop in temperature are detected. Shortening the cycle can also be driven by model performance metrics; when the model's prediction accuracy in climate-fluctuating regions drops below a preset threshold, iteration is automatically triggered.

[0084] Conversely, for climate-stable regions, the system extends the iteration cycle of the prediction model to reduce computational resource consumption. This model iteration strategy for climate-stable regions aims to optimize system resource utilization and reduce unnecessary computational overhead. In regions identified as climate-stable, the system reduces the iteration frequency of the condensation prediction model, extending it from the usual monthly iterations to semi-annual or annual iterations. This is based on the assumption that historical data patterns are relatively fixed under stable climates, allowing the model to maintain high accuracy without frequent adjustments. Extending the iteration cycle directly reduces the frequency of model training and parameter optimization by edge computing modules or cloud servers, thereby reducing CPU / GPU utilization, memory consumption, and network traffic, achieving effective savings in computational resources.

[0085] Through the above technical solution, this application can dynamically and adaptively adjust the operating time of the ventilation structure according to the temperature and humidity difference and its changing trend in the environment where the electricity metering box is located. This avoids the increased risk of condensation due to insufficient ventilation or the energy waste caused by excessive ventilation in the traditional fixed mode. In particular, the ventilation time is shortened when the temperature and humidity difference is decreasing and extended when it is increasing, making condensation control more accurate and energy-saving. At the same time, by dynamically adjusting the iteration cycle of the prediction model according to the regional climate stability, it ensures that the condensation prediction model maintains high accuracy and adaptability under different climatic conditions, avoiding misjudgments caused by model lag in areas with fluctuating climates and the waste of computing resources caused by frequent iterations in areas with stable climates. Overall, this solution improves the intelligence level and resource utilization efficiency of condensation control, effectively ensuring the long-term stable operation and metering accuracy of the electricity metering box.

[0086] In some of the embodiments described above in this application, a dual-engine recognition mode is proposed to identify anomalies and generate early warning information. In its implementation, rule judgment may be based only on threshold screening, which is easily affected by environmental interference and may lead to misjudgment. Trend analysis needs to be more comprehensively verified to eliminate random fluctuations, and the early warning information may lack detailed guidance, resulting in a high false alarm rate and low on-site handling efficiency.

[0087] In response, this application further proposes that in the aforementioned dynamic adaptive operation and maintenance and metering optimization method for electricity metering boxes, the dual-engine identification mode involves rule-based judgment to complete the initial screening of anomalies based on preset parameter thresholds, comparing the deviations of real-time parameters with preset safety thresholds and normal fluctuation ranges, and trend analysis to complete secondary verification by comparing historical data from the same period, long-term operating curves of the equipment, and data from similar equipment, thus eliminating misjudgments caused by accidental fluctuations and environmental interference. The combination of these two methods effectively reduces the false alarm rate of a single identification mode. The early warning information includes the time of anomaly occurrence, characteristic parameters, associated equipment number, installation location, and preliminary handling suggestions, providing clear guidance for on-site handling by operation and maintenance personnel.

[0088] Specifically, the dual-engine identification mode is an anomaly detection framework that combines two different analysis logics, aiming to improve the accuracy and robustness of identification through a complementary mechanism. This mode can be implemented by an edge computing module or a software module on a cloud server, processing data in parallel or serially to comprehensively determine whether an anomaly exists.

[0089] The rule-based judgment serves to quickly and initially screen for anomalies in real-time collected electrical, environmental, and enclosure status parameters. Its implementation can include: one approach is to preset a series of hard thresholds, such as upper voltage limits, lower current limits, and upper enclosure temperature limits; when real-time parameters exceed these fixed thresholds, a preliminary anomaly alarm is triggered. Another approach is to statistically determine the normal fluctuation range of each parameter based on historical operating data, and establish a confidence interval by calculating the mean and standard deviation; when real-time parameters deviate from this normal fluctuation range by a preset deviation percentage, it is determined as a potential anomaly. Furthermore, more complex preliminary judgments can be made by configuring logical rules, such as indicating potential electricity theft if the current remains zero and the power factor is abnormal.

[0090] The trend analysis serves to further verify potential anomalies initially identified by rule-based judgment, eliminating false alarms caused by accidental fluctuations, environmental interference, or momentary sensor malfunctions. This can be achieved through several methods: First, comparing current real-time data with historical data from the same period. For example, comparing the current hourly current value with the average current value for the same hour over the past few weeks or months. If the current value exceeds a threshold but follows a trend consistent with historical data from the same period, it may be a normal fluctuation. Second, analyzing the long-term operating curves of the equipment. For instance, constructing long-term trend lines for parameters using methods such as moving averages or exponential smoothing. If the current anomaly matches the long-term trend line, it may not be a sudden anomaly. Third, comparing the current equipment's operating data horizontally with the operating data of other equipment of the same type, batch, or region. If multiple devices simultaneously exhibit similar anomalies, it may be caused by external environmental factors rather than a single device malfunction. Furthermore, machine learning models, such as time series forecasting models, can be used to predict future parameter values, and the real-time values ​​can be compared with the predicted values. If the deviation is too large, it is considered an anomaly.

[0091] The purpose of the aforementioned early warning information is to provide maintenance personnel with a comprehensive and clear description of the abnormal situation and preliminary handling suggestions, thereby improving the efficiency and accuracy of on-site handling. This information can be generated by an edge computing module or a cloud server and sent to the maintenance terminal via SMS, app push notifications, email, etc. Its content may include: a timestamp of the abnormality occurrence, accurate to the second; the specific characteristic parameters that triggered the early warning and their current values, such as voltage phase A: 250V, exceeding the normal range; the unique associated device number of the metering box where the abnormality occurred; the precise installation location information of the metering box, such as geographical coordinates or a detailed address; and preliminary handling suggestions preset according to the type of abnormality, such as checking for loose wiring connections, verifying the electricity meter display data, and checking the sealing of the box.

[0092] Through the above technical solutions, this application effectively solves the problem of high false alarm rate in traditional single identification modes. Rule-based judgment can quickly filter out potential anomalies, avoiding the resource consumption of complex trend analysis on all data and improving identification efficiency. Trend analysis, based on rule-based judgment, introduces contextual information in the time and space dimensions. By comparing with historical data, long-term curves, and data from similar devices, it can effectively distinguish between genuine anomalies and accidental fluctuations or environmental interference, thereby significantly reducing the false alarm rate. The collaborative work of the two engines makes anomaly identification both fast and accurate. In addition, the early warning information includes key information such as the time of anomaly occurrence, characteristic parameters, associated device number, installation location, and preliminary handling suggestions, providing clear guidance for maintenance personnel. This enables them to quickly locate the problem, determine the nature of the anomaly, and take preliminary handling measures, improving the efficiency and accuracy of on-site handling and avoiding blind investigation and delayed processing due to incomplete information.

[0093] In some of the solutions mentioned above in this application, a data platform is proposed to build a shared data platform for both power supply and consumption parties, realize full-process data traceability and automatic generation of operation and maintenance analysis reports. In this process, an imperfect access control mechanism may lead to data security risks and information privacy leaks. The lack of targeted analysis in the report content affects the decision-making efficiency of regional operation and maintenance management. Furthermore, incomplete coverage of the full life cycle traceability links leads to ambiguity in problem tracing and difficulty in defining responsibilities.

[0094] In response, this application further proposes that the data platform supports hierarchical management of permissions for both power suppliers and users, assigning data viewing, operation, and export permissions through role-based allocation to ensure data security and information privacy; the operation and maintenance analysis report includes statistics on high-frequency problem areas, analysis of equipment failure patterns, evaluation of operation and maintenance efficiency, and strategy optimization suggestions, providing data support for regional operation and maintenance management; the full lifecycle traceability covers all stages of data collection, condensation control, anomaly warning, inspection and handling, and error compensation, establishing a data chain linking each stage to achieve data backtracking, problem tracing, and responsibility identification at any node.

[0095] Specifically, the hierarchical access control supported by the data platform refers to implementing different levels of access control for data and functions within the platform based on user identity, responsibilities, or business needs. Its purpose is to ensure that only authorized users can access specific data or perform specific operations, thereby preventing unauthorized data leakage, tampering, or abuse, and maintaining data security and user privacy. This hierarchical access control can be implemented in several ways. It can be based on role-based access control, where the system predefines various roles, including administrators, senior maintenance personnel, general maintenance personnel, power supplier representatives, and power consumer representatives. Each role is assigned specific data viewing and operation permissions, such as data tagging, status update, and export permissions. After logging in, users automatically obtain the corresponding permissions based on their assigned role. Alternatively, it can be based on attribute-based access control. In addition to roles, user attributes such as department, geographical region, and time period can be considered, as well as data attributes such as sensitivity level and data type, to dynamically determine access permissions. For example, only maintenance personnel in a specific region can view equipment data in that region, or only those within a specific time period can export data.

[0096] Role segmentation refers to dividing the user group into different role categories based on the different responsibilities and needs of the power supply and consumption parties in the operation and maintenance and metering optimization process of the electricity metering box. Its function is to provide a foundation for hierarchical permission management, making permission allocation more refined and reasonable, and avoiding excessive or insufficient permission granting. The roles of the power supply party and the power consumption party can be clearly defined. The power supply party role typically has broader permissions to view and operate operation and maintenance data, while the power consumption party role mainly focuses on its own electricity consumption data, anomaly warnings, and operation and maintenance progress. Furthermore, roles can be further refined; within the power supply party, they can be divided into data analysts, on-site operation and maintenance personnel, and system administrators. Data viewing, operation, and export permissions refer to the specific types of operations that can be performed on various types of data stored in the data platform, such as electrical parameters, environmental parameters, box status parameters, warning information, inspection records, error compensation data, and control logs. Its function is to achieve refined permission control, meet the different data utilization needs of different roles, and prevent sensitive data from being illegally obtained or misused. Viewing permissions allow users to browse data but not modify or download it; operation permissions allow users to make specific modifications to the data, such as confirming warnings or updating inspection status. Export permissions allow users to download data locally for further analysis or archiving. These permissions can be assigned to different roles independently or in combination. The granularity of permissions can be refined to the data field level; some roles may only be able to view the device's operating status but not specific electricity consumption data; or they may only be able to export summary reports but not raw detailed data.

[0097] An operations and maintenance (O&M) analysis report is a structured document or interactive interface that integrates various O&M-related data obtained from a data platform and performs in-depth analysis to provide a comprehensive understanding of the O&M status of electricity metering boxes. Its purpose is to provide O&M decision-makers with quantitative data to help them understand the current O&M status, identify potential problems, evaluate work effectiveness, and formulate improvement strategies. The report can be automatically generated periodically, such as weekly or monthly, and pushed to relevant personnel via email or system notifications. The report content is presented in a combination of charts, tables, and text descriptions, intuitively displaying the analysis results. Alternatively, a customizable report generation function can be provided, allowing users to select the time frame, analysis dimensions, and key indicators to generate personalized analysis reports.

[0098] High-frequency problem area statistics refer to the statistical analysis and ranking of the frequency of various anomalies, malfunctions, or events requiring maintenance intervention in electricity metering boxes across different geographical or management areas. Its purpose is to help maintenance personnel quickly identify areas with high maintenance pressure and concentrated problems, thereby prioritizing resource allocation, strengthening inspections, or implementing targeted upgrades. The system can automatically calculate the number and type distribution of abnormal events in each area based on early warning information, inspection records, and fault reports, generating heat maps or ranking lists to visually display high-problem areas. Equipment failure pattern analysis refers to identifying the inherent patterns and trends in the time, type, cause, and environmental conditions of failures in electricity metering boxes and their components through mining and analysis of historical failure data. Its purpose is to help predict potential failures, optimize preventative maintenance plans, extend equipment life, and reduce failure rates. Statistical methods, such as failure frequency analysis and mean time between failures (MTBF), and machine learning algorithms, such as association rule mining and time series analysis, can be used to discover failure patterns of specific equipment models under specific environments. Operation and maintenance (O&M) efficiency assessment refers to the quantitative evaluation of the timeliness, accuracy, and resource consumption of O&M personnel or teams in handling abnormal events, completing inspection tasks, and implementing error compensation. Its purpose is to measure O&M performance, identify efficiency bottlenecks, and provide a basis for optimizing O&M processes and resource allocation. Assessment indicators may include average fault response time, average fault repair time, inspection task completion rate, error compensation success rate, and unit equipment O&M cost. The system automatically collects relevant data and calculates the assessment results. Strategy optimization suggestions refer to specific action plans proposed by the system or expert system based on the problems and patterns revealed in the O&M analysis report, aimed at improving the O&M management and metering optimization strategies of electricity metering boxes. Its purpose is to provide actionable guidance to help O&M managers continuously improve their O&M capabilities. Based on statistics of high-frequency problem areas, it is recommended to increase the inspection frequency or upgrade equipment in specific areas. Based on equipment failure pattern analysis, it is recommended to adjust the preventive maintenance cycle of specific equipment or replace vulnerable components.

[0099] Full lifecycle traceability refers to the complete and continuous recording and association of data and events generated by an electricity metering box from data acquisition onwards, through all key operation and maintenance stages, including condensation control, anomaly warning, inspection and handling, and error compensation. Its function is to build a traceable data chain, ensuring that data or events at any stage can be accurately traced back to their source and context. This includes including all core business process data—data acquisition, condensation control, anomaly warning, inspection and handling, and error compensation—within the traceability scope, ensuring comprehensive transparency of the operation and maintenance process. Establishing a data chain linking each stage refers to logically connecting and integrating data generated by different operation and maintenance stages within the data platform through technical means, forming a complete and traceable data flow. Its function is to break down data silos, allowing data from one stage to easily query related data from its preceding and subsequent stages, thereby constructing a complete event context. In database design, foreign key associations and event ID associations can be used to associate warning information with corresponding inspection work orders, inspection work orders with on-site handling records, and error compensation records with environmental data before compensation. Ultimately, this enables data backtracking at any node, problem tracing, and responsibility determination. Data backtracking at any node means that users can trace back from any point in the data chain, such as an anomaly warning record, upwards to its originating data source, like a sensor reading at a specific moment, or downwards to its subsequent processing, such as corresponding inspection records and handling results. Problem tracing refers to the ability to accurately identify the root cause, time of occurrence, and involved equipment and environmental conditions through the data chain when a fault, metering dispute, or maintenance accident occurs. Responsibility determination refers to clarifying the responsibilities and actions of each relevant party during the occurrence or handling of a problem, based on information such as operation logs, personnel sign-ins, and handling results recorded in the data chain, thereby providing a basis for performance evaluation, incident analysis, and improvement.

[0100] Through the aforementioned technical solutions, this application effectively addresses the shortcomings of the data platform in terms of access control, report analysis, and full lifecycle traceability. The introduction of hierarchical access control and role-based classification ensures refined control over data access and operation permissions for both power suppliers and users, thereby improving data security and information privacy protection, and avoiding the risks of data leakage and misuse. The operation and maintenance analysis report, by integrating high-frequency problem area statistics, equipment failure pattern analysis, operation and maintenance efficiency assessment, and strategy optimization suggestions, provides multi-dimensional and in-depth data support for regional operation and maintenance management, making operation and maintenance decisions more scientific and accurate, and effectively improving operation and maintenance efficiency and management level. Furthermore, the full lifecycle traceability mechanism covers all key aspects such as data collection, condensation control, anomaly warning, inspection and handling, and error compensation, and establishes a data chain linking each aspect, allowing data at any node to be traced back. This enables clear problem tracing and responsibility identification, greatly enhancing the transparency and credibility of the operation and maintenance process. These improvements work together to enable the data platform to more securely and efficiently support dynamic adaptive operation and maintenance and metering optimization of electricity metering boxes, providing better services to both power suppliers and users.

[0101] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A dynamic adaptation type operation and maintenance and metering optimization method for an electric energy metering box, characterized in that, The process includes the following steps: relying on the existing sensing devices and data transmission channels of the power metering box, electrical parameters, environmental parameters and box status parameters are collected in real time. The raw data is preprocessed by noise reduction and normalization to construct a three-dimensional dataset of electrical-environment-status. The dataset is then stored locally and preliminarily analyzed by the edge computing module to reduce the redundancy of remote transmission. Based on the pre-processed environmental parameters, the probability of condensation is calculated by a preset prediction model. Combined with the difference between the environmental parameters inside and outside the box, adaptive control logic is triggered to link the existing ventilation structure to achieve condensation prevention and control. Simultaneously, control data is recorded to iteratively optimize model parameters. Anomalies are extracted from electrical parameters and enclosure status parameters, multi-dimensional feature vectors are constructed, and anomalies are identified using a dual-engine mode of rule judgment and trend analysis. Warning levels are divided according to severity, and warning information is generated and pushed to the operation and maintenance terminal and linked to historical data comparison to improve the accuracy of identification. By combining the 3D dataset, the metering box is comprehensively scored, the equipment status level is divided according to the score, and the corresponding inspection cycle and key points are matched to form a closed-loop operation and maintenance mechanism of scoring-scheduling-inspection-re-evaluation. The metering error benchmark values ​​for different temperature and humidity ranges are preset. The edge computing module compares the temperature and humidity inside the box with the benchmark range in real time, calculates the error compensation value, and corrects the metering parameters of the electricity meter through remote commands. The compensation accuracy is verified regularly to ensure metering accuracy. Establish a data sharing platform for both power supply and consumption parties to synchronize early warning information, inspection records, error compensation data, and control logs, enabling full-process data traceability and automatically generating operation and maintenance analysis reports to provide data support for strategy iteration.

2. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The electrical parameters include voltage, current, power factor and power data; the environmental parameters include internal temperature and humidity, external temperature and humidity and atmospheric humidity; and the enclosure status parameters include door magnetic switch status, vibration parameters and humidity difference at the sealing point. Data acquisition adopts a combination of timed and triggered modes. Electrical parameters and environmental parameters are acquired at timed intervals, while door magnetic switch status and humidity difference at the sealing point are acquired at triggered intervals. Vibration parameters are acquired through a combination of timed and triggered modes. The preprocessing process also includes removing sensor interference, filling in missing data, filtering out outliers, and verifying the data validity through a preset data validity threshold to ensure the integrity and reliability of the 3D dataset.

3. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The preset prediction model is based on the dew point temperature calculation principle. It combines the temperature inside the box and the humidity difference between inside and outside the box to build a multi-parameter correlation model. It is iteratively trained by combining historical condensation data, corresponding environmental parameters and control effects to adapt to different regional climate characteristics and installation scenarios. The adaptive control logic uses the probability of condensation and the difference between parameters inside and outside the chamber as the core triggering conditions. When the probability reaches the preset threshold and the temperature and humidity difference meets the control requirements, the ventilation structure is activated and the running time is dynamically adjusted according to the size of the difference. When there is no significant difference in parameters between the inside and outside of the chamber but the humidity inside the chamber exceeds the standard, an intermittent ventilation mode is adopted, and the ventilation structure is started and stopped in a cycle at preset intervals to balance the condensation control effect and energy consumption.

4. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The abnormal features include the parameter mutation rate, voltage distortion rate, current fluctuation range and power factor deviation of electrical parameters, as well as the abnormal opening time of the door magnet, vibration mutation amplitude, vibration duration and humidity difference mutation value of the enclosure status parameters. The warning levels are divided into at least Level 1 and Level 2. Level 1 anomalies correspond to serious problems that affect electricity safety and metering accuracy, such as suspected electricity theft, line overload, and short circuit hazards. Level 2 anomalies correspond to potential hazards such as poor line contact, aging enclosure seals, and minor vibration interference. Different levels of early warning information correspond to different push priorities and processing procedures. Level 1 abnormal information is pushed to the operation and maintenance terminal in real time and triggers an audio and visual reminder. Level 2 abnormal information is included in the operation and maintenance to-do list and pushed according to priority.

5. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The comprehensive status score adopts a weighted quantitative evaluation, assigning corresponding weights to the stability of measurement accuracy, environmental adaptability and physical status of the enclosure. Among them, the weight of measurement accuracy stability is the highest, and the weights of the other two items can be dynamically adjusted according to the application scenario. Based on the rating, the equipment is divided into three levels: excellent, qualified, and warning. The inspection cycle of excellent equipment is extended, the inspection cycle of qualified equipment is maintained at the regular cycle and the focus is on investigating key warning items, and the inspection cycle of warning equipment is shortened and on-site maintenance is prioritized. After the inspection is completed, the status score is updated based on the on-site testing data to form a closed-loop evaluation system.

6. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The metering error reference values ​​are preset according to the meter box type, covering common types such as single-phase meters, three-phase meters and meter boxes connected via current transformers. Each type corresponds to an error reference value for different temperature and humidity ranges. The benchmark values ​​are obtained by combining laboratory standard environment simulation tests with calibration based on actual on-site operating data to ensure accuracy and applicability; Error compensation does not change the hardware structure of the electricity meter. It is achieved by adjusting the built-in software metering coefficient through remote commands. After compensation, the accuracy is verified at fixed intervals. The deviation between the compensated data and the data of the standard metering equipment is compared, and the error benchmark value is dynamically corrected.

7. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 2, characterized in that: The existing data transmission channels include at least one of RS485, power line carrier, low-power wireless and NB-IoT. The adaptation method is selected according to the signal coverage capability of the installation scenario. For complex scenarios, a multi-channel redundant transmission design is adopted. The edge computing module is integrated into the existing terminal equipment of the metering box, without the need for additional hardware, and realizes local data processing and analysis through embedded algorithms; The local storage of 3D datasets adopts partition management, and is archived according to data type, acquisition time and importance. Data indexes are established to improve the efficiency of retrieval, comparison and traceability.

8. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 3, characterized in that: The ventilation structure includes a door ventilation grille and a built-in ventilation fan. In the adaptive control logic, the ventilation duration is positively correlated with the temperature and humidity difference, and is dynamically adjusted in combination with the changing trend of environmental parameters. When the temperature and humidity difference is decreasing, shorten the ventilation time; when it is increasing, extend the ventilation time. The prediction model iteration cycle is set according to the regional climate stability. The cycle is shortened in areas with large climate fluctuations to adapt to environmental changes, while the cycle is extended in areas with stable climates to reduce computational resource consumption.

9. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 4, characterized in that: In the dual-engine recognition mode, rule judgment completes the initial screening of anomalies based on preset parameter thresholds, and compares the deviation of real-time parameters with preset safety thresholds and normal fluctuation ranges. Trend analysis performs secondary verification by comparing historical data from the same period, long-term operating curves of the equipment, and data from similar equipment, eliminating misjudgments caused by accidental fluctuations and environmental interference. The combination of these two methods effectively reduces the false alarm rate of a single identification mode. The early warning information includes the time of the anomaly, characteristic parameters, associated device number, installation location, and preliminary handling suggestions, providing clear guidance for maintenance personnel to handle the situation on-site.

10. The dynamic adaptation operation and maintenance and metering optimization method for electricity metering boxes according to claim 1, characterized in that: The data platform supports hierarchical management of permissions for both power suppliers and users, assigning data viewing, operation, and export permissions through role division to ensure data security and information privacy. The operation and maintenance analysis report includes statistics on high-frequency problem areas, analysis of equipment failure patterns, evaluation of operation and maintenance efficiency, and suggestions for strategy optimization, providing data support for regional operation and maintenance management; The full lifecycle traceability covers all stages of data collection, condensation control, anomaly early warning, inspection and handling, and error compensation. It establishes a data chain linking each stage to enable data backtracking, problem tracing, and responsibility identification at any node.

Images