Intelligent electric meter remote monitoring method and system and intelligent electric meter
By constructing a remote monitoring system for smart meters, dynamic adjustment of data acquisition frequency and communication strategy was achieved, along with second-level anomaly diagnosis at the edge and cloud optimization. This solved the problems of high resource consumption and response latency in traditional meter systems, and improved the system's reliability and operational efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-20
- Publication Date
- 2026-04-10
AI Technical Summary
Traditional electricity meter data acquisition systems consume high communication network bandwidth and cloud storage resources, have difficulty balancing global monitoring with instantaneous detail capture in their data acquisition modes, suffer from severe response delays, and lack intelligent adjustment capabilities at the terminal, resulting in resource waste and untimely handling of abnormal events.
A remote monitoring system for smart meters is constructed, integrating smart meter terminals, edge computing nodes, and a cloud service platform. This system enables dynamic adaptive adjustment of data acquisition frequency and communication strategies. Lightweight AI models perform second-level anomaly diagnosis and priority transmission at the edge, while deep data analysis and predictive optimization are conducted in the cloud.
It improves the precision of data capture and the speed of anomaly response, reduces cloud computing pressure and communication bandwidth consumption, and achieves comprehensive optimization of system reliability, energy efficiency and operation and maintenance efficiency.
Smart Images
Figure CN121842544A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of power systems, and in particular to remote monitoring methods, systems and smart meters for smart meters. Background Technology
[0002] With the deepening of smart grid construction and the rise of the energy internet concept, refined management and intelligent operation and maintenance of power systems have become an inevitable trend. As a key data node connecting the power grid and users, smart meters play an important role in ensuring power supply security, improving operational efficiency, combating electricity theft, and achieving demand-side response by monitoring their operation status and conducting in-depth analysis of electricity consumption data.
[0003] Traditional electricity meter data acquisition systems mostly adopt a centralized architecture, where electricity meter terminals collect data at fixed intervals and upload it directly to a central data center via a communication network. This model faces significant challenges in practical applications: First, the massive amount of data generated by the huge number of electricity meters puts enormous pressure on communication network bandwidth and cloud storage and computing resources, and the transmission cost is high. Secondly, the fixed-frequency acquisition mode is difficult to meet the needs of both global monitoring and instantaneous detail capture. It may miss important instantaneous fault characteristics during critical periods such as peak electricity consumption, while generating a large amount of redundant data during off-peak hours at night, resulting in a waste of resources. Finally, the long chain from data collection to cloud processing and alarm generation results in significant delays in responding to abnormal events such as electricity theft and equipment malfunctions, failing to meet real-time requirements. Furthermore, the meter terminal itself lacks intelligent adjustment capabilities, making it susceptible to data loss during network outages and unable to activate self-protection mechanisms under abnormal operating conditions. Summary of the Invention
[0004] To address the aforementioned issues, this application provides a method, system, and smart meter for remote monitoring of smart meters.
[0005] The remote monitoring method, system, and smart meter provided in this application adopt the following technical solution: A remote monitoring system for smart meters includes a meter body, which integrates a smart meter terminal. The output of the smart meter terminal is communicatively connected to an edge computing node. The output of the edge computing node is communicatively connected to a cloud service platform. The output of the cloud service platform is communicatively connected to a user terminal. The smart meter terminal is used to collect electricity consumption data and generate raw data packets. The edge computing node is used to receive the raw data packets and preprocess them to generate standard data frames. The preprocessing includes data cleaning, format normalization, and preliminary anomaly diagnosis. The cloud service platform is used to receive and store the standard data frames and perform deep data analysis to generate monitoring results. The user terminal is used to receive and display the monitoring results.
[0006] As a preferred technical solution of this application, the smart meter terminal includes a data acquisition module, a local storage module, a first communication module, and a terminal management module. The data acquisition module is used to acquire voltage, current, power, and power quality parameters. The local storage module is used to cache raw data packets. The first communication module is used to establish a communication link with the edge computing node. The terminal management module is used to adjust the acquisition frequency of the data acquisition module or the communication strategy of the first communication module according to instructions received from the edge computing node or the cloud service platform.
[0007] As a preferred technical solution of this application, the edge computing node includes a second communication module, an edge computing engine, and a policy execution module. The second communication module is used to communicate with multiple smart meter terminals and the cloud service platform. The edge computing engine is used to run a preprocessing process. The policy execution module is used to issue control policies to one or more designated smart meter terminals based on the results of preliminary anomaly diagnosis or instructions received from the cloud service platform. Preliminary anomaly diagnosis includes: real-time analysis of the original data packets based on preset thresholds or adaptive learning models, and generating anomaly markers to be appended to standard data frames when anomalies are detected; When generating the standard data frame, the edge computing engine assigns different transmission priorities to different standard data frames based on the data characteristics of the original data packet or the data importance tag issued by the cloud service platform. The second communication module then sends the standard data frame to the cloud service platform using a differentiated communication protocol based on the transmission priority.
[0008] As a preferred technical solution of this application, the cloud service platform includes a data warehouse, an analysis engine, and an alarm and feedback module. The data warehouse is used to store historical and real-time standard data frames, the analysis engine is used to perform deep data analysis, and the alarm and feedback module is used to generate alarm information and push it to the relevant user terminal and the policy execution module when the analysis result of the analysis engine meets the preset alarm conditions.
[0009] As a preferred technical solution of this application, based on historical data stored in the data warehouse, the analysis engine obtains a power consumption pattern benchmark through machine learning model training. The analysis engine identifies deviations by comparing the received real-time standard data frames with the power consumption pattern benchmark. The alarm information generated by the alarm and feedback module is at least partially based on the identified deviations.
[0010] As a preferred technical solution of this application, the monitoring results display provided by the user terminal includes a visual dashboard and an interactive interface. The visual dashboard is used to display real-time electricity consumption data, historical trend comparisons, and alarm information lists. The interactive interface is used by users to set alarm parameter thresholds, query data details for a specific time period, and manually issue control commands to a specified smart meter terminal.
[0011] A remote monitoring method for smart meters, based on the aforementioned remote monitoring system for smart meters, includes the following steps: Step 1: The smart meter terminal collects electricity consumption data and generates raw data packets; Step two: The edge computing node receives the raw data packets and preprocesses them to generate standard data frames; Step 3: The cloud service platform receives standard data frames and performs in-depth data analysis to generate monitoring results; Step four: The monitoring results are received and displayed by the user terminal.
[0012] As a preferred technical solution of this application, in step two, more specifically, the original data packet is first cleaned and formatted, and then the preprocessed data is analyzed in real time based on preset rules or a lightweight machine learning model to perform preliminary anomaly diagnosis. Finally, a standard data frame containing anomaly markers is generated based on the diagnosis results, and a control strategy for a specific smart meter terminal is generated and issued based on the diagnosis results or instructions from the cloud service platform.
[0013] As a preferred technical solution of this application, in step three, more specifically, the received standard data frame is first compared and analyzed with the historical power consumption pattern benchmark. When the analysis result indicates that there is power consumption abnormality, equipment failure risk or suspected power theft, alarm information is generated and pushed. Then, based on the long-term data analysis results, the historical power consumption pattern benchmark and the rules or model parameters sent to the edge computing nodes for preliminary anomaly diagnosis are optimized or updated.
[0014] In summary, this application includes at least the following beneficial technical effects of the remote monitoring method, system, and smart meter for smart meters: This application constructs a closed-loop system integrating smart meter terminals, edge computing nodes, cloud service platforms, and user terminals. It achieves dynamic adaptive adjustment of data acquisition frequency and communication strategy, second-level anomaly diagnosis and priority transmission based on a lightweight AI model on the edge side, and predictive deep analysis with continuous optimization in the cloud. This improves the precision of data capture and the speed of anomaly response while reducing cloud computing pressure and communication bandwidth consumption, ultimately achieving comprehensive optimization of system reliability, energy efficiency, and operation and maintenance efficiency. Attached Figure Description
[0015] Figure 1 This is a schematic diagram of the system architecture of this application; Figure 2 This is a flowchart of the remote monitoring method of this application. Detailed Implementation
[0016] The following is in conjunction with the appendix Figure 1-2 This application will be described in further detail.
[0017] See Figure 1-2 The smart meter remote monitoring system includes a meter body, a smart meter terminal integrated within the meter body, an edge computing node connected to the output of the smart meter terminal, a cloud service platform connected to the output of the edge computing node, and a user terminal connected to the output of the cloud service platform. The smart meter terminal is used to collect electricity consumption data and generate raw data packets. The edge computing node is used to receive the raw data packets and preprocess them to generate standard data frames. The preprocessing includes data cleaning, format normalization, and preliminary anomaly diagnosis. The cloud service platform is used to receive and store the standard data frames and perform in-depth data analysis to generate monitoring results. The user terminal is used to receive and display the monitoring results. The smart meter terminal includes a data acquisition module, a local storage module, a first communication module, and a terminal management module. The data acquisition module is used to collect voltage, current, power, and power quality parameters. The local storage module is used to cache raw data packets. The first communication module is used to establish a communication link with the edge computing node. The terminal management module is used to adjust the acquisition frequency of the data acquisition module or the communication strategy of the first communication module according to the instructions received from the edge computing node or cloud service platform.
[0018] In this application, the terminal management module receives instructions from the cloud service platform to increase the frequency of the data acquisition module from once per minute to once every 10 seconds during peak electricity consumption periods to capture details of instantaneous power fluctuations, while reducing it to once every 5 minutes during off-peak hours at night to save energy. The first communication module adopts a dynamic communication strategy: when the edge computing node reports network congestion, it automatically switches to the LoRaWAN protocol for low-power, long-distance transmission of basic data; when the network is unobstructed, it uses 4G / 5G networks to transmit complete data packets at high speed. The local storage module adopts a ring buffer design, which can cache 72 hours of raw data when communication is interrupted and retransmit data according to priority after the connection is restored. The terminal also integrates a temperature sensor. When an abnormal temperature of the meter body is detected, the terminal management module will automatically trigger a protection mechanism, suspend high-frequency acquisition, and send a device fault warning to the upper level. This application improves the precision of data acquisition and system reliability by dynamically adjusting the acquisition and communication strategies.
[0019] The edge computing node includes a second communication module, an edge computing engine, and a policy execution module. The second communication module communicates with multiple smart meter terminals and a cloud service platform. The edge computing engine runs a preprocessing flow. The policy execution module issues control policies to one or more designated smart meter terminals based on the results of preliminary anomaly diagnosis or instructions received from the cloud service platform. Preliminary anomaly diagnosis includes: real-time analysis of raw data packets based on preset thresholds or adaptive learning models, and generating anomaly tags to be appended to standard data frames when anomalies are detected. When generating standard data frames, the edge computing engine assigns different transmission priorities to different standard data frames based on the data characteristics of the raw data packets or the data importance tags issued by the cloud service platform. The second communication module sends the standard data frames to the cloud service platform using differentiated communication protocols according to the transmission priorities. In this application, the edge computing engine deploys a lightweight LSTM autoencoder model to stream data packets received from multiple smart meter terminals. First, data cleaning is performed, such as filtering current value abrupt changes caused by electromagnetic interference and filling in missing values through linear interpolation. During the format normalization stage, heterogeneous data is unified into IEEE C37.118 standard frames. In the preliminary anomaly diagnosis stage, the model learns the baseline of each meter's power consumption pattern in real time. When a meter is detected to have a continuous power load during non-production periods, a suspected electricity theft marker is automatically added to the standard data frame. The policy execution module acts synchronously: on the one hand, it sends an instantaneous power dual-frequency acquisition control policy to the meter to collect evidence; on the other hand, through priority management of the second communication module, data frames marked as urgent are uploaded to the cloud platform via a 5G URLLC link, while regular data is transmitted in batches via NB-IoT. This application reduces cloud computing pressure and achieves second-level anomaly response through edge-side adaptive learning and multi-priority transmission. The lightweight LSTM autoencoder deployed in the edge computing engine has the following structure: The input layer receives a sequence of power values from a single electricity meter at 120 consecutive time points. The sequence is encoded by three layers of LSTM units, with 64, 32, and 16 neurons in each layer, respectively, to achieve a "lightweight" design. The model is trained with historical normal data, and the loss function is the mean squared error. The reconstruction error is minimized by the Adam optimizer. During anomaly diagnosis, the model calculates the reconstruction error of the real-time power sequence. When the error continuously exceeds the dynamic threshold set according to the 95th quantile of historical data, it is judged as an anomaly and a label is attached. When the second communication module continuously monitors the cloud platform and determines that the end-to-end latency is greater than 500 milliseconds or the packet loss rate is greater than 5% for 30 seconds, it will be considered as "network congestion". At this time, it will send a command to the electricity meter terminal to switch to LoRaWAN. If the standard data frame carries suspected electricity theft or equipment failure mark, or is marked as critical equipment data by the cloud platform, it will be defined as emergency priority and uploaded through 5G eMBB or URLLC link. Data containing only preliminary abnormality mark is defined as high priority and transmitted through 4G / 5G ordinary channel. Regular data is classified as ordinary priority and uploaded in batches using NB-IoT.
[0020] The cloud service platform includes a data warehouse, an analysis engine, and an alarm and feedback module. The data warehouse stores historical and real-time standard data frames, the analysis engine performs deep data analysis, and the alarm and feedback module generates alarm information and pushes it to relevant user terminals and policy execution modules when the analysis results of the analysis engine meet preset alarm conditions. Based on historical data stored in the data warehouse, the analysis engine obtains a power consumption pattern benchmark through machine learning model training. The analysis engine identifies deviations by comparing the received real-time standard data frames with the power consumption pattern benchmark. The alarm information generated by the alarm and feedback module is at least partially based on the identified deviations. In this application, the data warehouse stores electricity consumption data, and the analysis engine uses the Transformer architecture to train the electricity consumption pattern benchmark. First, the historical data is clustered to identify the electricity consumption curve, and then a dynamic benchmark is generated by combining external features. When real-time standard data frames flow in, the engine performs multi-dimensional comparisons. The alarm and feedback module adopts a cascading response: emergency alarms are pushed to the user terminals of operation and maintenance personnel through the Kafka message queue, and update instructions are issued to edge nodes to lower the voltage sag detection threshold of the relevant area meters. The cloud service platform retrains the benchmark model every week to improve the accuracy of anomaly detection. In this application, predictive maintenance is achieved through the continuously optimized benchmark model and real-time feedback closed loop.
[0021] The monitoring results provided by the user terminal include a visual dashboard and an interactive interface. The visual dashboard displays real-time electricity consumption data, historical trend comparisons, and alarm information lists. The interactive interface allows users to set alarm parameter thresholds, query data details for specific time periods, and manually issue control commands to designated smart meter terminals.
[0022] The remote monitoring method for smart meters includes the following steps: Step 1: The smart meter terminal collects electricity consumption data and generates raw data packets; Step two involves the edge computing node receiving the raw data packets and preprocessing them to generate standard data frames. More specifically, in step two, the raw data packets are first cleaned and formatted. Then, the preprocessed data is analyzed in real time based on preset rules or a lightweight machine learning model to perform preliminary anomaly diagnosis. Finally, a standard data frame containing anomaly markers is generated based on the diagnosis results, and a control strategy for a specific smart meter terminal is generated and issued based on the diagnosis results or instructions from the cloud service platform. Step 3: The cloud service platform receives standard data frames and performs in-depth data analysis to generate monitoring results. More specifically, in step 3, the received standard data frames are first compared and analyzed with historical power consumption patterns. When the analysis results indicate that there is abnormal power consumption, equipment failure risk, or suspected power theft, alarm information is generated and pushed. Then, based on long-term data analysis results, the historical power consumption pattern benchmark and the rules or model parameters sent to the edge computing nodes for preliminary anomaly diagnosis are optimized or updated. Step four: The user terminal receives and displays the monitoring results; First, in step one, the data acquisition module of the smart meter terminal collects quality parameters such as voltage, current, and power according to a preset strategy, generates raw data packets, and temporarily stores them in a local circular buffer. Then, in step two, the edge computing node receives the data through the second communication module, and its edge computing engine immediately executes a preprocessing procedure: first, data cleaning and format regularization are performed; then, a lightweight LSTM autoencoder model is run for preliminary anomaly diagnosis, generating standard data frames with anomaly markers. Simultaneously, the strategy execution module issues control strategies to specific meters based on the diagnostic results or cloud platform instructions. Finally, in step three, the cloud service platform receives the standard data frames and stores them in the data warehouse. The library's analysis engine performs in-depth comparative analysis of real-time data with dynamically optimized historical electricity consumption patterns. When deviations are identified, the alarm and feedback module generates cascading alarms via a Kafka message queue and pushes them to the user terminal. Simultaneously, update instructions are sent to edge nodes, and the baseline model is retrained weekly to achieve predictive maintenance. Finally, in step four, the user terminal's visual dashboard and interactive interface display monitoring results, alarm lists, and historical trends in real time, and support users to manually query data or issue control commands. This forms a closed-loop, adaptive, and efficient remote monitoring process from data acquisition, edge preprocessing, cloud-based intelligent analysis to result display.
[0023] This application constructs a closed-loop system integrating smart meter terminals, edge computing nodes, cloud service platforms, and user terminals. It achieves dynamic adaptive adjustment of data acquisition frequency and communication strategy, second-level anomaly diagnosis and priority transmission based on a lightweight AI model on the edge side, and predictive deep analysis with continuous optimization in the cloud. This improves the precision of data capture and the speed of anomaly response while reducing cloud computing pressure and communication bandwidth consumption, ultimately achieving comprehensive optimization of system reliability, energy efficiency, and operation and maintenance efficiency.
[0024] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.
Claims
1. A remote monitoring system for smart meters, characterized in that: The device includes an electricity meter body, which integrates a smart meter terminal. The output of the smart meter terminal is communicatively connected to an edge computing node. The output of the edge computing node is communicatively connected to a cloud service platform. The output of the cloud service platform is communicatively connected to a user terminal. The smart meter terminal is used to collect electricity consumption data and generate raw data packets. The edge computing node is used to receive the raw data packets and preprocess them to generate standard data frames. The preprocessing includes data cleaning, format normalization, and preliminary anomaly diagnosis. The cloud service platform is used to receive and store the standard data frames and perform deep data analysis to generate monitoring results. The user terminal is used to receive and display the monitoring results.
2. The remote monitoring system for smart meters according to claim 1, characterized in that: The smart meter terminal includes a data acquisition module, a local storage module, a first communication module, and a terminal management module. The data acquisition module is used to acquire voltage, current, power, and power quality parameters. The local storage module is used to cache raw data packets. The first communication module is used to establish a communication link with the edge computing node. The terminal management module is used to adjust the acquisition frequency of the data acquisition module or the communication strategy of the first communication module according to instructions received from the edge computing node or the cloud service platform.
3. The remote monitoring system for smart meters according to claim 2, characterized in that: The edge computing node includes a second communication module, an edge computing engine, and a policy execution module. The second communication module is used to communicate with multiple smart meter terminals and the cloud service platform. The edge computing engine is used to run a preprocessing process. The policy execution module is used to issue control policies to one or more designated smart meter terminals based on the results of preliminary anomaly diagnosis or instructions received from the cloud service platform. Preliminary anomaly diagnosis includes: real-time analysis of the original data packets based on preset thresholds or adaptive learning models, and generating anomaly markers to be appended to standard data frames when anomalies are detected; When generating the standard data frame, the edge computing engine assigns different transmission priorities to different standard data frames based on the data characteristics of the original data packet or the data importance tag issued by the cloud service platform. The second communication module then sends the standard data frame to the cloud service platform using a differentiated communication protocol based on the transmission priority.
4. The remote monitoring system for smart meters according to claim 3, characterized in that: The cloud service platform includes a data warehouse, an analysis engine, and an alarm and feedback module. The data warehouse is used to store historical and real-time standard data frames. The analysis engine is used to perform deep data analysis. The alarm and feedback module is used to generate alarm information and push it to the relevant user terminal and the policy execution module when the analysis results of the analysis engine meet the preset alarm conditions.
5. The remote monitoring system for smart meters according to claim 4, characterized in that: Based on historical data stored in the data warehouse, the analysis engine obtains a power consumption pattern benchmark through machine learning model training. The analysis engine identifies deviations by comparing the received real-time standard data frames with the power consumption pattern benchmark. The alarm information generated by the alarm and feedback module is at least partially based on the identified deviations.
6. The remote monitoring system for smart meters according to claim 5, characterized in that: The monitoring results provided by the user terminal include a visual dashboard and an interactive interface. The visual dashboard is used to display real-time electricity consumption data, historical trend comparisons, and alarm information lists. The interactive interface is used by users to set alarm parameter thresholds, query data details for specific time periods, and manually issue control commands to designated smart meter terminals.
7. A method for remote monitoring of smart meters, the smart meter remote monitoring system according to claim 6, characterized in that: Includes the following steps: Step 1: The smart meter terminal collects electricity consumption data and generates raw data packets; Step two: The edge computing node receives the raw data packets and preprocesses them to generate standard data frames; Step 3: The cloud service platform receives standard data frames and performs in-depth data analysis to generate monitoring results; Step four: The monitoring results are received and displayed by the user terminal.
8. The remote monitoring method for smart meters according to claim 7, characterized in that: In step two, more specifically, the original data packets are first cleaned and formatted. Then, the preprocessed data is analyzed in real time based on preset rules or a lightweight machine learning model to perform preliminary anomaly diagnosis. Finally, a standard data frame containing anomaly markers is generated based on the diagnosis results, and a control strategy for a specific smart meter terminal is generated and issued based on the diagnosis results or instructions from the cloud service platform.
9. The remote monitoring method for smart meters according to claim 8, characterized in that: In step three, more specifically, the received standard data frames are first compared and analyzed with historical power consumption patterns. When the analysis results indicate that there is an abnormal power consumption, equipment failure risk, or suspected power theft, alarm information is generated and pushed. Then, based on long-term data analysis results, the historical power consumption pattern benchmark and the rules or model parameters sent to the edge computing nodes for preliminary anomaly diagnosis are optimized or updated.