Search

CN-121995302-A - Electric energy meter online self-detection method and system based on edge calculation

CN121995302ACN 121995302 ACN121995302 ACN 121995302ACN-121995302-A

Abstract

The invention discloses an online self-detection method and system of an electric energy meter based on edge calculation, which belong to the field of data processing, wherein the online self-detection method of the electric energy meter based on the edge calculation comprises the following steps of constructing an electricity consumption behavior feature library at an edge calculation node based on historical electricity consumption data, establishing a baseline prediction model through time sequence analysis, and generating a theoretical electricity consumption prediction value; the method has the beneficial effects that the lightweight baseline prediction model is deployed at the edge calculation node to perform primary screening, so that the data processing capacity and the network transmission pressure of the cloud are effectively reduced, external factors such as environmental factors and population variation are introduced into the cloud platform to perform secondary verification on primary anomalies, the accuracy of anomaly detection is improved, and the advantage of low time delay of edge calculation is brought into play through the cooperative work of the edge and the cloud, and the powerful calculation capacity of the cloud is utilized.

Inventors

  • CAO JUN
  • ZHANG ZHANG
  • LIU YEQING

Assignees

  • 霍立克电气有限公司

Dates

Publication Date
20260508
Application Date
20260330

Claims (10)

  1. 1. The electric energy meter online self-detection method based on the edge calculation is characterized by comprising the following steps of: Acquiring power consumption data acquired by the electric energy meter in real time, comparing the power consumption data with the theoretical power consumption predicted value to calculate a difference value, judging that the electric energy meter is normal in operation if the difference value does not exceed a first threshold value, marking the electric energy meter as a preliminary abnormal state if the difference value exceeds the first threshold value, and uploading preliminary abnormal information and associated data to a cloud platform; Receiving preliminary abnormal information and associated data uploaded by an edge computing node through a cloud platform, calling an external data interface to obtain environmental factors influencing electricity utilization, analyzing historical contemporaneous behavior characteristics of a user, identifying whether personnel variation exists in a current time period, dynamically correcting a baseline prediction model according to the identified environmental factors influencing electricity utilization and personnel variation information, and generating a corrected theoretical electricity utilization predicted value; And (3) carrying out secondary comparison on the corrected theoretical electricity consumption predicted value and the real-time electricity consumption data, judging that the electric energy meter is normal if the difference value falls back to the first threshold value range, issuing a command for restoring the normal state to the edge computing node, and marking the electric energy meter as a final abnormal state if the difference value still exceeds the first threshold value, generating alarm information and notifying operation and maintenance personnel.
  2. 2. The method for online self-detection of the electric energy meter based on edge calculation according to claim 1, wherein the method is characterized in that the electric energy meter on-line self-detection method comprises the steps of constructing an electric behavior feature library based on historical electric energy data at an edge calculation node, establishing a baseline prediction model through time sequence analysis to generate a theoretical electric energy prediction value, acquiring the electric energy meter on-line collected electric energy data, comparing the electric energy meter on-line collected electric energy data with the theoretical electric energy prediction value to calculate a difference value, judging that the electric energy meter is normal in operation if the difference value does not exceed a first threshold value, marking the electric energy meter as a preliminary abnormal state if the difference value exceeds the first threshold value, and uploading preliminary abnormal information and associated data to a cloud platform, and the method specifically comprises the following steps: At the edge computing node, receiving power consumption original data uploaded by the electric energy meter in real time through a communication interface, extracting target parameters after carrying out integrity check and format conversion on the power consumption original data, and storing the target parameters into a local cache as historical power consumption data; The method comprises the steps of calling historical electricity consumption data, calculating the mean value, variance and variation trend of electricity consumption in each period by adopting a sliding window statistical method, and constructing an electricity consumption behavior feature library, wherein the characteristics of the electricity consumption behavior feature library comprise daily electricity consumption modes, holiday effects and season fluctuation; The method comprises the steps of acquiring power consumption data acquired by the electric energy meter in real time, comparing the power consumption data with a theoretical power consumption predicted value, calculating deviation percentage, judging that the electric energy meter is normal in operation if the deviation is within a first threshold value, continuing monitoring, marking the electric energy meter as a preliminary abnormal state if the deviation exceeds the first threshold value, and packaging and uploading preliminary abnormal information and associated data to a cloud platform.
  3. 3. The method for online self-detection of electric energy meter based on edge calculation according to claim 1, wherein the step of receiving preliminary abnormal information and associated data uploaded by edge calculation nodes through a cloud platform, calling an external data interface to obtain environmental factors affecting electricity consumption, analyzing historical contemporaneous behavior characteristics of a user, identifying whether personnel variation exists in a current time period, dynamically correcting a baseline prediction model according to the identified environmental factors affecting electricity consumption and personnel variation information, and generating corrected theoretical electricity consumption prediction values specifically comprises the following steps: Receiving preliminary abnormal information and associated data uploaded by an edge computing node through a cloud platform, calling an API interface to obtain a third party data source, and judging whether an environment factor influencing electricity consumption exists in an abnormal period or not, wherein the environment factor influencing electricity consumption comprises a temperature factor and an electricity price factor; analyzing the historical contemporaneous electricity consumption behavior of the user, identifying a typical electricity consumption mode of the user through a clustering algorithm, comparing the similarity between an electricity consumption curve of an abnormal period and the historical mode, and comprehensively judging whether personnel variation exists or not by combining external data; If the environmental factors and personnel fluctuation affecting electricity consumption are identified, a baseline prediction model is dynamically corrected according to the fluctuation type, a corrected theoretical electricity consumption predicted value is generated, the corrected theoretical electricity consumption predicted value is secondarily compared with electricity consumption data acquired in real time, corrected deviation is calculated, and meanwhile deviation change trend before and after correction is compared to be used as a judgment basis.
  4. 4. The method for online self-detection of an electric energy meter based on edge calculation according to claim 3, wherein the corrected theoretical electricity consumption predicted value is calculated by the following steps: define the original theoretical electricity consumption predicted value as The corrected theoretical electricity consumption predicted value is T refers to time; Temperature influencing factor Wherein Is the actual temperature at the time t, As a reference to the temperature of the liquid, Is a temperature sensitivity coefficient; Electricity price influencing factor Wherein The electricity price change amount at the time t is, As the reference electricity price, the utility model has the advantages of, Is the price elastic coefficient; factor of influence of the number of people Wherein The change amount of the number of people at the time t, As a basis for the number of people in a living, If a factor does not exist, the corresponding factor takes 1, and the corrected theoretical electricity utilization predicted value 。
  5. 5. The method for online self-detection of electric energy meter based on edge calculation according to claim 1, wherein the steps of secondarily comparing the corrected theoretical electricity consumption predicted value with real-time electricity consumption data, judging that the electric energy meter is normal if the difference value falls back to a first threshold value range, and issuing an instruction for restoring the normal state to an edge calculation node, and marking the electric energy meter as a final abnormal state if the difference value still exceeds the first threshold value, generating alarm information and notifying operation and maintenance personnel, specifically comprises: Performing secondary comparison on the corrected theoretical electricity consumption predicted value and real-time electricity consumption data, if the corrected deviation falls back to a preset first threshold range, judging that the primary abnormal state is caused by normal electricity consumption fluctuation, and sending an instruction for restoring the normal state to an edge computing node when the electric energy meter actually runs normally; if the corrected deviation still exceeds the first threshold value, finally judging that the electric energy meter has faults or anomalies, marking the electric energy meter as a final anomaly state, generating alarm information comprising anomaly types, anomaly grades and recommended treatment measures, and pushing the alarm information to an operation and maintenance personnel terminal through a short message or APP; And storing all data of the abnormal event into an abnormal database for subsequent baseline prediction model optimization and fault analysis.
  6. 6. The edge-calculation-based on-line self-test method for an electric energy meter according to any one of claims 1 to 5, further comprising: The cloud platform counts the number of the preliminary abnormal information reported by a plurality of edge computing nodes in the same geographic area in real time in the same period, if the number exceeds a preset second threshold value, the occurrence of an area power grid event is judged, and a pause uploading instruction is issued to all the edge computing nodes in the area; Monitoring whether regional power grid event confirmation information from a power grid dispatching system or an external data source exists or not through a cloud platform within a preset time length third threshold, if the confirmation information is received before the waiting time length reaches the third threshold, marking that all relevant anomalies in the time period are caused by regional power grid events in the cloud platform, not checking for the second time, issuing an instruction for clearing data locally cached when the regional power grid event occurs to an edge computing node, and if the confirmation information is not received yet when the third threshold is over, issuing a restoration uploading instruction to the edge computing node through the cloud platform, and after the cloud platform receives the supplementary transmission data, sequentially executing the second time checking and final judgment.
  7. 7. An online self-detection system of an electric energy meter based on edge calculation, which is characterized by comprising: The primary abnormality judgment module is used for constructing an electricity consumption behavior feature library based on historical electricity consumption data at an edge computing node, establishing a baseline prediction model through time sequence analysis to generate a theoretical electricity consumption prediction value, acquiring the electricity consumption data acquired by the electric energy meter in real time, comparing the electricity consumption data with the theoretical electricity consumption prediction value to calculate a difference value, judging that the electric energy meter operates normally if the difference value does not exceed a first threshold value, marking the electric energy meter as a primary abnormal state if the difference value exceeds the first threshold value, and uploading primary abnormal information and associated data to the cloud platform; The dynamic correction module is used for receiving the preliminary abnormal information and the associated data uploaded by the edge computing node through the cloud platform, calling an external data interface to acquire environmental factors influencing electricity utilization, analyzing historical contemporaneous behavior characteristics of a user, identifying whether personnel variation exists in the current time period, dynamically correcting the baseline prediction model according to the identified environmental factors influencing electricity utilization and personnel variation information, and generating a corrected theoretical electricity utilization predicted value; The secondary comparison judging module is used for carrying out secondary comparison on the corrected theoretical electricity consumption predicted value and the real-time electricity consumption data, judging that the electric energy meter is normal if the difference value falls back to the first threshold range, sending an instruction for restoring the normal state to the edge computing node, marking the electric energy meter as a final abnormal state if the difference value still exceeds the first threshold, generating alarm information and notifying operation and maintenance personnel.
  8. 8. The edge-calculation-based electric energy meter online self-detection system according to claim 7, wherein the preliminary abnormality judgment module comprises: the data processing unit is used for receiving the power consumption original data uploaded by the electric energy meter in real time at the edge computing node through the communication interface, extracting target parameters after carrying out integrity check and format conversion on the power consumption original data, and storing the target parameters into a local cache to serve as historical power consumption data; The system comprises a theoretical value acquisition unit, a power consumption behavior feature library, a time sequence prediction algorithm, a dynamic first threshold value and a power consumption prediction unit, wherein the theoretical value acquisition unit is used for acquiring historical power consumption data, calculating the mean value, variance and variation trend of the power consumption in each period by adopting a sliding window statistical method, and constructing the power consumption behavior feature library; The method comprises the steps of selecting an uploading unit, comparing power consumption data acquired by the electric energy meter in real time with a theoretical power consumption predicted value, calculating a deviation percentage, judging that the electric energy meter is normal in operation if the deviation is within a first threshold value, continuing monitoring, marking the electric energy meter as a preliminary abnormal state if the deviation exceeds the first threshold value, and packing and uploading preliminary abnormal information and associated data to a cloud platform.
  9. 9. The edge-computing-based power meter online self-detection system of claim 7, wherein the dynamic correction module comprises: The environment factor judging unit is used for receiving the preliminary abnormal information and the associated data uploaded by the edge computing node through the cloud platform, calling the API interface to obtain a third party data source, and judging whether the environment factor influencing the electricity consumption exists in the abnormal period or not, wherein the environment factor influencing the electricity consumption comprises a temperature factor and an electricity price factor; The personnel variation judging unit is used for analyzing the historical contemporaneous electricity consumption behavior of the user, identifying the typical electricity consumption mode of the user through a clustering algorithm, comparing the similarity between the electricity consumption curve of the abnormal period and the historical mode, and comprehensively judging whether personnel variation exists or not by combining external data; The prediction value correction unit is used for dynamically correcting the baseline prediction model according to the change type to generate a corrected theoretical power consumption prediction value if the environmental factors and personnel changes affecting power consumption are identified, carrying out secondary comparison on the corrected theoretical power consumption prediction value and power consumption data acquired in real time, calculating corrected deviation, and simultaneously comparing deviation change trend before and after correction to be used as a judgment basis.
  10. 10. The electric energy meter online self-detection system based on edge calculation according to claim 9, wherein the corrected theoretical electricity consumption predicted value is calculated by the following method: define the original theoretical electricity consumption predicted value as The corrected theoretical electricity consumption predicted value is T refers to time; Temperature influencing factor Wherein Is the actual temperature at the time t, As a reference to the temperature of the liquid, Is a temperature sensitivity coefficient; Electricity price influencing factor Wherein The electricity price change amount at the time t is, As the reference electricity price, the utility model has the advantages of, Is the price elastic coefficient; factor of influence of the number of people Wherein The change amount of the number of people at the time t, As a basis for the number of people in a living, If a factor does not exist, the corresponding factor takes 1, and the corrected theoretical electricity utilization predicted value 。

Description

Electric energy meter online self-detection method and system based on edge calculation Technical Field The invention belongs to the field of data processing, and particularly relates to an electric energy meter online self-detection method and system based on edge calculation. Background At present, the monitoring of the running state of the electric energy meter generally depends on an electricity consumption information acquisition system, and the intelligent electric energy meter is used as an edge node to be responsible for acquiring and uploading data. In the prior art, the common practice is that the electric energy meter directly uploads mass raw data to the cloud master station according to fixed frequency, and the cloud platform is used for centralized storage, processing and analysis. However, with the development of smart grids, the number of electric energy meters has proliferated, and this mode of transmitting all data to a cloud platform for processing faces a great challenge. On one hand, the cloud needs to process a large amount of redundant normal data, so that the calculation resource waste and the network bandwidth pressure are increased, and the system load is increased, and on the other hand, when the electric energy meter is abnormal, the data are ‌ transmitted to the cloud at a fixed frequency to perform post analysis ‌, so that the problems are difficult to find and position in time. In addition, the traditional anomaly detection method is often judged only based on a fixed threshold value, dynamic changes of electricity consumption behaviors of users are not fully considered, such as population fluctuation caused by factors such as holiday and tour, temporary resident increase in home and the like, natural fluctuation of electricity consumption is caused, if the electricity consumption is directly judged to be abnormal without distinction, false alarm is easily generated, and user experience and operation and maintenance efficiency are affected. To sum up, the existing electric energy meter is not intelligent enough in abnormality detection and needs improvement. Disclosure of Invention Based on this, it is necessary to provide an online self-detection method and system for an electric energy meter based on edge calculation aiming at the above problems. The embodiment of the invention is realized in such a way that the electric energy meter online self-detection method based on edge calculation comprises the following steps: The method comprises the steps of constructing a power consumption behavior feature library based on historical power consumption data at an edge computing node, establishing a baseline prediction model through (lightweight) time sequence analysis to generate a theoretical power consumption predicted value, acquiring power consumption data acquired by an electric energy meter in real time, comparing the power consumption data with the theoretical power consumption predicted value to calculate a difference value, judging that the electric energy meter is normal in operation if the difference value does not exceed a first threshold value, marking the electric energy meter as a preliminary abnormal state if the difference value exceeds the first threshold value, and uploading preliminary abnormal information and associated data to a cloud platform; Receiving preliminary abnormal information and associated data uploaded by an edge computing node through a cloud platform, calling an external data interface to obtain environmental factors influencing electricity utilization, analyzing historical contemporaneous behavior characteristics of a user, identifying whether personnel variation exists in a current time period, dynamically correcting a baseline prediction model according to the identified environmental factors influencing electricity utilization and personnel variation information, and generating a corrected theoretical electricity utilization predicted value; And (3) carrying out secondary comparison on the corrected theoretical electricity consumption predicted value and the real-time electricity consumption data, judging that the electric energy meter is normal if the difference value falls back to the first threshold value range, issuing a command for restoring the normal state to the edge computing node, and marking the electric energy meter as a final abnormal state if the difference value still exceeds the first threshold value, generating alarm information and notifying operation and maintenance personnel. In one embodiment, the invention provides an online self-detection method of an electric energy meter based on edge calculation, which comprises the steps of constructing an electricity consumption behavior feature library based on historical electricity consumption data at an edge calculation node, establishing a baseline prediction model through time sequence analysis to generate a theoretical electricity consumption prediction value, acquiring the