Electricity consumption metering method combined with power user grouping

Abstract

The present application provides an electricity metering method combined with power user grouping, which relates to the field of electricity metering technology, including: querying the historical electricity consumption records of corporate users in the region, grouping users, and determining multiple user clusters; dividing multiple user clusters into daily electricity consumption intervals, optimizing electricity metering strategies based on the division results, and determining multiple optimized metering schemes; establishing an electricity metering anomaly detection plug-in; executing electricity metering for corporate users according to multiple optimized metering schemes, synchronously monitoring electricity metering anomalies through the electricity metering anomaly detection plug-in, and dynamically adjusting multiple optimized metering schemes based on the anomaly detection results. This application can solve the technical problem in the prior art that electricity metering fails to effectively adapt to the dynamic changes in electricity consumption behavior, resulting in insufficient metering efficiency and accuracy. By combining user grouping, more accurate and efficient electricity metering is achieved, which improves the pertinence and efficiency of metering.

Description

Technical Field

[0001] The present application relates to the technical field of electricity metering, and in particular to an electricity meter electricity metering method combined with power user grouping. Background Art

[0002] Electricity metering refers to the process of accurately recording and calculating the amount of electricity consumed by a user over a specific period of time. Electricity user segmentation involves categorizing users into different groups based on their electricity usage characteristics, behavioral patterns, and lifestyle habits. This helps power companies better understand their electricity needs, enabling personalized electricity management and services. In electricity metering, segmentation can be used to implement differentiated pricing policies, such as peak and off-peak pricing, or to develop energy-saving incentives for different user groups, encouraging users to use electricity during periods of low grid load, thereby optimizing overall grid efficiency and energy allocation. Currently, most electricity metering methods are based on fixed peak and off-peak timeframes. However, electricity usage is dynamic and can vary with time of day, season, operating mode, and other factors. While current metering methods already account for peak and off-peak timeframes and some electricity usage characteristics, their adaptability to complex or dynamic electricity usage patterns remains limited, resulting in insufficient metering efficiency and accuracy. Especially when faced with diverse electricity consumption patterns and dynamically changing user needs, static metering strategies cannot effectively adapt. Differences in behavior among different users can easily cause metering deviations, affecting the accuracy and efficiency of overall metering.

[0003] In summary, the existing technology has technical problems such as the lack of dynamic adaptability of electricity metering and the mismatch between metering strategies and dynamic electricity consumption behaviors, which will cause deviations between metering results and actual electricity consumption, resulting in insufficient metering efficiency and accuracy. Summary of the Invention

[0004] The purpose of this application is to provide an electricity metering method combined with electricity user grouping to solve the technical problems in the existing technology that electricity metering lacks dynamic adaptability, the mismatch between metering strategies and dynamic electricity consumption behaviors, and the deviation between metering results and actual electricity consumption, resulting in insufficient metering efficiency and accuracy.

[0005] In view of the above problems, the present application provides an electric meter electricity consumption metering method combined with electric power user grouping, wherein the electric meter electricity consumption metering method combined with electric power user grouping includes: querying the historical electricity consumption records of corporate users in the area, grouping users according to the historical electricity consumption records, and determining multiple user clusters; dividing the multiple user clusters into daily electricity consumption intervals, performing electricity metering strategy optimization based on the division results, and determining multiple optimized metering schemes; establishing an electricity metering anomaly detection plug-in based on the multiple user clusters; executing electricity metering for corporate users according to the multiple optimized metering schemes, synchronously monitoring electricity metering anomalies through the electricity metering anomaly detection plug-in, and performing dynamic adjustment of the multiple optimized metering schemes according to the anomaly detection results.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By querying the historical electricity consumption records of enterprise users in the area, users are grouped according to the historical electricity consumption records to determine multiple user clusters; daily electricity consumption intervals are divided for each of the multiple user clusters, and electricity metering strategy optimization is performed based on the division results to determine multiple optimized metering schemes; an electricity metering anomaly detection plug-in based on the multiple user clusters is established; electricity metering of enterprise users is performed according to the multiple optimized metering schemes, electricity metering anomalies are synchronously monitored through the electricity metering anomaly detection plug-in, and dynamic adjustment of the multiple optimized metering schemes is performed based on the anomaly detection results. In other words, by grouping, users are classified according to similar electricity consumption behavior patterns, and the electricity consumption data of each user cluster is further analyzed, and their daily electricity consumption is divided into different intervals. The metering strategy is customized to achieve more refined management, and abnormal electricity consumption is identified according to the historical behavior patterns of the user clusters. The metering strategy is dynamically adjusted to achieve dynamic adaptability of electricity metering, so that the metering scheme can be flexibly adjusted according to the real-time behavior of the user and the detected abnormal changes, avoiding the accumulation of long-term deviations, achieving more accurate and efficient metering of electricity consumption, and improving the pertinence and efficiency of metering.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0010] Figure 1 This is a flow chart of the electricity metering method for this application combined with electricity user grouping.

[0011] Figure 2 This is a flow chart of determining an optimized metering solution in the electricity metering method based on electricity user grouping in this application. DETAILED DESCRIPTION

[0012] This application solves the technical problems in the prior art that the lack of dynamic adaptability in electricity metering and the mismatch between metering strategies and dynamic electricity consumption behaviors will cause deviations between metering results and actual electricity consumption, resulting in insufficient metering efficiency and accuracy, by providing an electricity metering method that combines electricity user clustering. By clustering, users are classified according to similar electricity consumption behavior patterns, and the electricity consumption data of each user cluster is further analyzed, and their daily electricity consumption is divided into different intervals. The metering strategy is customized to achieve more refined management, and abnormal electricity consumption is identified based on the historical behavior patterns of the user cluster. The metering strategy is dynamically adjusted to achieve dynamic adaptability of electricity metering, so that the metering scheme can be flexibly adjusted according to the user's real-time behavior and detected abnormal changes, avoiding the accumulation of long-term deviations, achieving more accurate and efficient metering of electricity consumption, and improving the pertinence and efficiency of metering.

[0013] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0014] For examples, please see the attached Figure 1 The present application provides a method for measuring electricity consumption by an electric meter in combination with grouping of electric power users, wherein the method for measuring electricity consumption by an electric meter in combination with grouping of electric power users specifically comprises the following steps:

[0015] Step 1: Query the historical electricity consumption records of enterprise users in the area, group the users according to the historical electricity consumption records, and determine multiple user clusters.

[0016] Specifically, the power company's database or data management system is queried for historical electricity usage records of corporate users in the region, including information such as electricity consumption, usage time, and peak usage over the past period. Monthly, weekly, and daily electricity usage feature sets are then collected from these historical records for multiple corporate users. Each feature set includes peak and valley curves, power factor curves, load curves, and load fluctuation curves, but the collection frequency varies. For example, monthly features are collected every six hours, weekly features every half hour, and daily features every minute. High-frequency features are extracted from each of these sets, and the most representative monthly, weekly, and daily data are selected. Using a clustering algorithm, the monthly, weekly, and daily data are used to group the corporate users three times, one at a time, based on their electricity usage behavior and needs, ensuring that users within each group have similar electricity usage behaviors. Finally, multiple user clusters are obtained. By extracting and clustering features at the monthly, weekly, and daily levels, we can more comprehensively understand users' electricity usage behavior and ensure that users are accurately classified.

[0017] Step 2: Divide the multiple user clusters into daily electricity consumption intervals, perform electricity metering strategy optimization according to the division results, and determine multiple optimized metering schemes.

[0018] Specifically, daily electricity usage characteristic data is obtained for each user cluster, including peak and valley curves, power factor curves, load curves, and fluctuation curves. Each user cluster is then segmented into daily electricity usage intervals, with the day divided into multiple time intervals according to a preset time step, typically a short one. The electricity usage data for each time interval is fitted, and the centroid of that interval is calculated. This is the middle value of each segmented curve, reflecting the electricity usage characteristics (such as peak and valley values, load fluctuation, etc.) during that time period. The centroid data is used to cluster the time intervals, assigning intervals with similar electricity usage characteristics to the same cluster. Clustering generates multiple time thresholds and multiple sets of electricity usage characteristic centroids, representing the demarcation points of different electricity usage patterns. These sets clearly distinguish different electricity usage characteristic intervals and form a time-of-day electricity usage pattern map. The electricity usage characteristic centroid sets represent the electricity usage characteristics of each cluster center. Based on these multiple sets of electricity usage characteristic centroids, the electricity metering strategy is optimized, including adjusting the sampling frequency and data sampling accuracy, resulting in multiple optimized metering strategies. A first optimized metering scheme is constructed based on multiple time thresholds and multiple optimized metering strategies, allowing the metering strategy to adapt to the electricity usage characteristics of each time period. A dispersion analysis is performed on the electricity usage characteristic set of each user cluster for the first week. If the dispersion is less than or equal to a predetermined discrete scalar, the first optimized metering scheme is added to the multiple optimized metering schemes. Conversely, the overall difference between each user and other users is analyzed, and the first optimized metering scheme is compensated before being added to the multiple optimized metering schemes. By dividing daily electricity usage into intervals and optimizing metering strategies, the electricity metering scheme for each user cluster can be personalized to adapt to its unique electricity usage characteristics, effectively improving the targetedness and efficiency of metering and ensuring accurate electricity usage data in different time periods.

[0019] Step 3: Establish an electricity meter anomaly detection plug-in based on the multiple user clusters.

[0020] Specifically, monthly, weekly, and daily electricity usage information for each user cluster is obtained, including peak and valley power consumption, power factor, load fluctuation, and voltage and current fluctuations. Based on daily electricity usage characteristics and historical data, the user cluster's electricity usage patterns and changing trends are analyzed. Key electricity metering anomaly factors are defined based on the user cluster's electricity usage characteristics, including but not limited to current exceeding the standard, voltage deviation, energy deviation, and power factor anomaly. Statistical analysis and threshold setting are performed at different time scales (such as monthly, weekly, and daily), with anomaly factor thresholds defined for each user cluster to accommodate different electricity demand and patterns. A separate detection branch is established for each user cluster's electricity metering anomaly factor, comprehensively considering all anomaly factors for that user cluster. Each user cluster is mapped and integrated with the corresponding electricity metering anomaly detection branch to create an electricity metering anomaly detection plug-in. The electricity metering anomaly detection plug-in manages the monitoring logic and alarm mechanism for each user cluster, dynamically adapting to the electricity usage behavior of different user clusters and performing anomaly detection based on actual data. Furthermore, the electricity metering anomaly detection plug-in should be able to access data from devices such as smart meters to enable real-time monitoring. When an anomaly is detected, the meter anomaly detection plug-in quickly issues an alert, notifying relevant management personnel. Based on long-term monitoring data, the anomaly factor thresholds and detection rules are regularly updated to adapt to changes in power usage patterns within user clusters. By building a cluster-based meter anomaly detection plug-in, we can centrally monitor power usage anomalies across each user cluster, quickly respond, and issue alerts.

[0021] Step 4: Execute electricity metering of enterprise users according to the multiple optimized metering schemes, synchronously monitor electricity metering anomalies through the electricity metering anomaly detection plug-in, and perform dynamic adjustment of the multiple optimized metering schemes according to the anomaly detection results.

[0022] Specifically, the electricity consumption of enterprise users is measured in real time according to multiple optimized metering schemes, including real-time data collection, data storage, and regular reporting. According to multiple optimized metering schemes, parameters such as current, voltage, and power are collected and measured according to the set sampling frequency and accuracy. At the same time, the electricity meter anomaly detection plug-in is started to monitor electricity consumption data in real time and analyze it according to preset electricity meter anomaly factors (such as current exceeding the standard, voltage deviation, etc.). The electricity meter anomaly detection plug-in continuously receives real-time data streams and compares them with the set thresholds to determine whether an anomaly has occurred. When an anomaly is detected, the electricity meter anomaly detection plug-in records the occurrence of the abnormal event, including information such as the type of anomaly, the time of occurrence, and duration, and immediately feedbacks these abnormal information to notify relevant management personnel for subsequent processing.

[0023] Based on the anomaly detection results of the electricity metering anomaly detection plug-in, analyze the abnormal electricity metering frequency and type of each user cluster, and evaluate its impact on the electricity metering plan. Dynamically adjust multiple optimized metering plans based on multiple abnormal electricity metering frequencies. If the abnormal electricity metering frequency is low, fine-tune the metering plan for individual users, adjust the compensation coefficient or optimize the metering parameters; if the abnormal electricity metering frequency is high, it is necessary to analyze and adjust the overall metering plan, re-evaluate the electricity consumption characteristics and update the metering plan. After the adjustment, continue to monitor electricity consumption and continuously optimize the metering plan based on the new data to form a closed-loop management. Through real-time anomaly detection and dynamic adjustment mechanism, it is possible to respond to electricity anomalies in a timely manner, and the electricity metering of corporate users can be more accurate and adapt to the actual electricity needs of different users.

[0024] Furthermore, step one of this application includes:

[0025] Based on the historical electricity consumption records, monthly electricity consumption feature sets, weekly electricity consumption feature sets and daily electricity consumption feature sets are collected for multiple corporate users, wherein the electricity consumption features include electricity peak and valley curves, power factor curves, electricity load curves and load fluctuation curves; high-frequency features are extracted from the monthly electricity consumption feature sets, weekly electricity consumption feature sets and daily electricity consumption feature sets respectively to determine monthly electricity consumption information, weekly electricity consumption information and daily electricity consumption information; using the K-means algorithm, the multiple corporate users are clustered once according to the monthly electricity consumption information to obtain a primary clustering result; the primary clustering result is clustered twice according to the weekly electricity consumption information to obtain a secondary clustering result; the secondary clustering result is clustered three times according to the daily electricity consumption information to obtain the multiple user clusters.

[0026] Specifically, historical electricity consumption records are collected through smart meters, power company databases, or other power detection equipment, including data such as electricity consumption, electricity consumption time, and power. Based on historical electricity consumption records, monthly, weekly, and daily electricity consumption feature sets are collected for multiple corporate users, and electricity consumption feature sets at different time scales are constructed, including electricity peak and valley curves, power factor curves, electricity load curves, and load fluctuation curves. Each curve reflects the user's electricity consumption behavior and patterns at different time scales. The electricity peak and valley curve describes the changes in electricity consumption peaks and valleys, reflecting the user's peak electricity consumption period and its fluctuations; the power factor curve is used to evaluate the effective utilization of electricity and can reflect the quality and efficiency of electricity consumption; the electricity load curve shows the changes in electricity load over time and analyzes the overall trend of electricity consumption; the load fluctuation curve describes the fluctuation of load within a certain period of time and identifies the stability of electricity consumption. Each power consumption characteristic curve is collected at a different frequency. For example, monthly power consumption characteristic sets are collected every six hours to analyze long-term power consumption trends and patterns within a month; weekly power consumption characteristic sets are collected every half hour to analyze medium-term power consumption patterns and changes within a week; and daily power consumption characteristic sets are collected every minute to analyze short-term power consumption behavior and fluctuations within a day. Enterprise users typically refer to different businesses within a specific area (such as an industrial park or commercial district). Each enterprise, as an independent power consumption entity, may have different power consumption behaviors and patterns depending on factors such as industry type, production scale, and operating hours.

[0027] High-frequency features are extracted from the monthly, weekly, and daily electricity usage feature sets. Multiple similarity comparisons and center calculations are performed to ensure the representativeness and validity of the selected features. Multiple similar high-frequency features are aggregated into a comprehensive curve, making the results smoother and easier to understand. Finally, monthly, weekly, and daily electricity usage information is determined. The K-means algorithm is used to perform a preliminary segmentation of multiple enterprise users based on their monthly electricity usage information, forming initial user clusters. The K-means algorithm is an iterative cluster analysis algorithm that selects K initial center points, either randomly or based on some heuristic method. For each data point in the dataset, its distance to each center point is calculated and assigned to the cluster represented by the closest center point. The center point of each cluster is updated by calculating the mean of all data points within the cluster to obtain a new center point. This process is repeated until a stopping condition is met, such as the change in the center point being less than a threshold or the maximum number of iterations being reached. By clustering based on monthly electricity usage information, the differences between users can be preliminarily divided, and users are grouped according to the similarities in their monthly electricity usage information, so that users with similar electricity usage behaviors are gathered into the same cluster.

[0028] Based on the primary clustering, similarly, user groups are further refined based on weekly electricity usage information to capture the characteristics and variations of weekly electricity usage, making each user cluster more representative. Finally, the secondary clustering results are subjected to a tertiary clustering based on daily electricity usage information to obtain the final user clusters. The purpose of the tertiary clustering is to capture subtle differences in daily electricity usage variations, ensuring that electricity usage behavior within each user cluster is as similar as possible and that electricity usage behavior outside of each cluster is as different as possible. Through hierarchical clustering, users can be accurately divided into different groups, allowing for personalized metering and management strategies to be implemented for different groups. By extracting and clustering features at the monthly, weekly, and daily levels, a more comprehensive understanding of user electricity usage behavior can be achieved, ensuring that users are accurately classified. Different metering strategies can be formulated for different users based on their characteristics, making metering more accurate and targeted.

[0029] Furthermore, the present application further comprises the following steps:

[0030] Randomly select the electricity consumption features of the first month, where the electricity consumption features of the first month include the electricity peak-valley curve of the first month, the power factor curve of the first month, the electricity load curve of the first month, and the load fluctuation curve of the first month; perform a similarity comparison between the electricity peak-valley curve of the first month and the electricity peak-valley curves of other months in the monthly electricity consumption feature set, and count the amount of data with a similarity greater than a similarity threshold, which is set as the first centrality; if the first centrality is greater than a predetermined scalar, perform a similarity comparison between the power factor curve of the first month and the power factor curves of other months in the monthly electricity consumption feature set, and count the amount of data with a similarity greater than the similarity threshold, which is set as the second centrality; if the second centrality is greater than the predetermined scalar, calculate and obtain the third centrality of the electricity load curve of the first month; if the third centrality is greater than the predetermined scalar, calculate and obtain the fourth centrality of the load fluctuation curve of the first month; if the fourth centrality is greater than the predetermined scalar, add the electricity consumption features of the first month to the high-frequency monthly electricity consumption feature set, and perform curve fitting on the high-frequency monthly electricity consumption feature set to obtain the monthly electricity consumption information.

[0031] Specifically, a monthly electricity consumption feature set randomly selects electricity consumption characteristics from a month as the first monthly electricity consumption feature set. These characteristics include the first month's peak-valley curve, the first month's power factor curve, the first month's load curve, and the first month's load fluctuation curve. These curves respectively demonstrate the peak and valley levels of electricity consumption, power efficiency, overall load, and monthly load fluctuations. The first month's peak-valley curve is then compared with the peak-valley curves of other months in the monthly electricity consumption feature set to calculate similarity. Similarity comparison is a data analysis method used to assess the degree of similarity between two or more curves. A pre-set similarity threshold is used to determine whether two curves are sufficiently similar. The number of data points with similarities exceeding the threshold is counted, i.e., the number of other monthly electricity consumption peak-valley curves that are similar to the first month's peak-valley curve is calculated as the first centrality. The first centrality reflects the degree of similarity between the peak-valley curves of that month and those of other months. If a high proportion of similarities are achieved, the peak-valley curve of that month is representative.

[0032] The predetermined scalar is a pre-set threshold used to determine whether the representativeness of the peak-valley curve of electricity consumption in the first month is high enough. If the first centrality exceeds the predetermined scalar, it means that the peak-valley curve of electricity consumption in the first month is highly representative in the data set, so it is necessary to further analyze the similarity of the power factor curve of the first month. Similar to the similarity comparison of the general electricity peak-valley curve, the power factor curve of the first month is compared with the power factor curves of other months to calculate the similarity. The amount of data with a similarity greater than the similarity threshold is counted, that is, how many other power factor curves are similar to the power factor curve of the first month. This number is the second centrality. When the second centrality is greater than the predetermined scalar, it means that the power factor curve of the first month is also highly representative.

[0033] Similarly, the first month's electricity load curve is compared with the other months' electricity load curves for similarity, and the third centrality is calculated. If the third centrality is greater than a predetermined scalar, the first month's electricity load curve is also highly representative. The first month's load fluctuation curve is then calculated to obtain the fourth centrality. This series of conditional judgments and calculations ensures that each feature undergoes rigorous similarity analysis, thereby ensuring that the final selected high-frequency features have a higher degree of credibility. If the fourth centrality is greater than a predetermined scalar, the first month's electricity consumption features are highly representative, and they are added to the high-frequency monthly electricity consumption feature set. Curve fitting is performed on the high-frequency monthly electricity consumption feature set, fitting identical high-frequency feature curves into a composite curve to obtain comprehensive monthly electricity consumption information. Curve fitting is a mathematical method used to find a mathematical function that best describes a set of data points. Through multiple similarity comparisons and centrality calculations, the final selected features are ensured to be highly representative and effective. Curve fitting aggregates multiple similar high-frequency features into a composite curve, resulting in a smoother and easier-to-understand result, which helps identify the electricity usage patterns of enterprise users.

[0034] Further, as attached Figure 2 As shown, step 2 of this application includes:

[0035] A first user cluster is randomly selected from the multiple user clusters, and first-day electricity consumption information of the first user cluster is obtained, wherein the first-day electricity consumption information includes a first-day electricity peak-valley fitting curve, a first-day power factor fitting curve, a first-day electricity load fitting curve, and a first-day load fluctuation fitting curve; a day is divided according to a predetermined time step to determine multiple time intervals, and multiple peak-valley fitting segment centroids, multiple factor fitting segment centroids, multiple load fitting segment centroids, and multiple fluctuation fitting segment centroids are obtained for the multiple time intervals; the multiple time intervals are clustered based on the multiple peak-valley fitting segment centroids, multiple factor fitting segment centroids, multiple load fitting segment centroids, and multiple fluctuation fitting segment centroids to determine multiple time thresholds and multiple electricity consumption feature centroid sets; electricity metering strategy optimization configuration is performed based on the multiple electricity consumption feature centroid sets to determine multiple optimized metering strategies, wherein the electricity metering strategies include sampling frequency and data sampling accuracy; a first optimized metering scheme is constructed based on the multiple time thresholds and the multiple optimized metering strategies, and is added to the multiple optimized metering schemes.

[0036] Specifically, a user cluster is randomly selected from multiple user clusters as the first user cluster, and the electricity consumption information of the first user cluster on a certain day, that is, the electricity consumption information of the first day, is obtained, including the peak and valley fitting curve of electricity consumption on the first day, the power factor fitting curve on the first day, the electricity load fitting curve on the first day, and the load fluctuation fitting curve on the first day. According to the preset time step, one day is divided into multiple time intervals. Usually this predetermined time step is short, such as once every 5 minutes. For each time interval, the segmented centroid of the peak and valley fitting curve, the power factor fitting curve, the electricity load fitting curve, and the load fluctuation fitting curve is calculated. The centroid is usually the statistical center point of the segmented curve, such as the mean or median, representing the electricity consumption characteristics of the interval.

[0037] Using multiple peak-valley fitting segment centroids, multiple factor fitting segment centroids, multiple load fitting segment centroids, and multiple fluctuation fitting segment centroids, a clustering algorithm (such as K-means) is used to cluster multiple time intervals. Similar time intervals are grouped together to determine multiple time thresholds and multiple sets of power consumption feature centroids. Based on the similarity of the segment centroids, time intervals with identical or similar power consumption characteristics are identified and clustered into a single cluster, representing time periods within a day with similar power consumption characteristics. This cluster is then used to further identify turning points (time thresholds) and typical features of power consumption patterns. Significant differences in the centroid values ​​of different time intervals result in different clusters, with the boundary between each cluster forming a time threshold. The central centroid of each cluster represents the typical power consumption feature value for that cluster, and these centroids form a set of power consumption feature centroids.

[0038] The electricity metering strategy is optimized based on multiple sets of electricity consumption feature centroids, including adjusting the sampling frequency and data sampling accuracy. For example, during peak hours and periods of large load fluctuations, it is recommended to set a higher sampling frequency to accurately capture electricity consumption fluctuations and ensure real-time data. During off-peak hours or periods of relatively stable load, the sampling frequency can be appropriately reduced to reduce data storage and transmission costs. During periods of peak electricity consumption or large power factor fluctuations, a higher sampling accuracy is set to ensure the accuracy of the acquired data information and facilitate subsequent analysis and anomaly detection. During periods of stable load, the sampling accuracy can be reduced to balance the requirements of data volume and accuracy. By setting different sampling frequency and accuracy requirements for each time period, multiple different optimized metering strategies are formed, allowing the metering solution to flexibly respond to changes in users' electricity consumption patterns.

[0039] Based on the multiple time thresholds and optimized metering strategies determined, the time thresholds and metering strategies are combined to form a highly targeted and detailed first optimized metering plan, including sampling frequency, accuracy, and other configurations within different time intervals. The first week's electricity usage characteristics of the first user cluster are obtained and analyzed for dispersion to assess the stability of the electricity usage characteristics within the user cluster. If the dispersion is small (i.e., the user cluster's electricity usage behavior is similar), the user cluster is considered to have consistent electricity usage patterns, and the first optimized metering plan can be directly added to the multiple optimized metering plans. If the dispersion is large, indicating significant electricity usage variability within the user cluster, the overall difference between each user in the first user cluster and other users is analyzed. If the difference exceeds the tolerance range, a weekly electricity consumption compensation coefficient is calculated based on the user electricity consumption variability deviation to balance users with large differences. After compensating the first optimized metering plan based on the weekly electricity consumption compensation coefficient, it is added to the multiple optimized metering plans, including sampling frequency and accuracy settings for different time periods, to meet the personalized electricity usage characteristics of the user cluster. By refining the time period division of daily electricity consumption characteristics within user clusters, clustering based on centroids and optimizing metering strategy configuration, the metering solution can be more accurately adapted to the user's electricity consumption behavior, thereby improving the efficiency and accuracy of the metering system.

[0040] Furthermore, the present application further comprises the following steps:

[0041] The multiple peak-valley fitting segment centroids are arranged in chronological order according to the time interval to determine a peak-valley segment centroid sequence, and the peak-valley segment centroid sequence is clustered to determine a first time threshold set; a second time threshold set, a third time threshold set and a fourth time threshold set are obtained by clustering the multiple factor fitting segment centroids, the multiple load fitting segment centroids and the multiple fluctuation fitting segment centroids; a union operation is performed on the first time threshold set, the second time threshold set, the third time threshold set and the fourth time threshold set in the threshold order to obtain the multiple time thresholds; wherein, the peak-valley segment centroid sequence is clustered to determine the first time threshold set, including: if the deviation of two adjacent peak-valley fitting segment centroids is less than or equal to a predetermined deviation threshold, the time intervals corresponding to the two adjacent peak-valley fitting segment centroids are merged, and the merged time interval corresponds to the average of the two adjacent peak-valley fitting segment centroids; if the deviation of two adjacent peak-valley fitting segment centroids is greater than the predetermined deviation threshold, a time dividing line is set, and the iterative merging is continued.

[0042] Specifically, multiple peak-valley segment centroids are arranged in chronological order by time interval, forming an ordered sequence of centroids that reflects the temporal trends of peak-valley electricity usage. This sequence of peak-valley segment centroids is then clustered to identify time periods with similar electricity usage characteristics. During the clustering process, the sequence of peak-valley segment centroids is traversed, and the deviations between two adjacent peak-valley segment centroids are compared. A predetermined deviation threshold is a pre-set critical value that determines whether two centroids are sufficiently close to allow the corresponding time intervals to be merged. If the deviation between two adjacent peak-valley segment centroids is less than or equal to the predetermined deviation threshold, the peak-valley electricity usage behavior of these two time intervals is considered similar. Therefore, the time intervals corresponding to these two peak-valley segment centroids are merged into a larger time interval, corresponding to the mean of the two adjacent peak-valley segment centroids. If the deviation between two adjacent peak-valley segment centroids is greater than the predetermined deviation threshold, the two time intervals should be considered different. A time divider is then set between the centroids, and the iterative merging process continues.

[0043] For example, assume the following peak-valley segmented centroid sequence (in kilowatt-hours): 100, 120, 140, 150, 152, 200; a preset deviation threshold of 10. Comparing 100 and 120, if the deviation is 20 greater than the threshold, a time split is set; comparing 120 and 140, if the deviation is 20 greater than the threshold, a time split is set; comparing 140 and 150, if the deviation is 10 equal to the threshold, the values ​​are merged into 145; comparing 150 and 152, if the deviation is 2 less than the threshold, the values ​​are merged into 151; comparing 152 and 200, if the deviation is 18 greater than the threshold, a time split is set. Based on the centroid deviation division results, the boundaries of the user's peak, valley, and transitional periods within a day are obtained, thereby forming a first time threshold set.

[0044] Similarly, multiple factor-fitting segment centroids, multiple load-fitting segment centroids, and multiple fluctuation-fitting segment centroids are independently clustered to obtain the second, third, and fourth time threshold sets. The first, second, third, and fourth time threshold sets are then unioned in the order of the thresholds. A union operation combines different time threshold sets (e.g., the first, second, third, and fourth time threshold sets) to form a comprehensive threshold set encompassing all time intervals. The core of the union operation is to integrate the time thresholds from each set without duplication, ensuring that the resulting time thresholds fully reflect the dynamic trends of user electricity consumption. Time nodes from each time threshold set are merged into a comprehensive time series in chronological order. Thresholds with the same order are merged; that is, if two or more thresholds are adjacent or overlapping in time, they are merged into a wider time interval. According to the changes in actual electricity consumption data, the time dividing line is flexibly set to ensure the true reflection of electricity consumption characteristics. Through the deviation analysis of adjacent centroids, the clustering results are ensured to have high accuracy to avoid the influence of noise on the clustering effect. The time thresholds from multiple sources are merged to form a comprehensive time threshold set.

[0045] Furthermore, the present application further comprises the following steps:

[0046] Obtain a first-week electricity consumption feature set for the first user cluster, where the first-week electricity consumption feature set includes a first-week electricity consumption peak-valley curve set, a first-week power factor curve set, a first-week electricity consumption load curve set, and a first-week load fluctuation curve set; perform a comprehensive dispersion calculation based on the first-week electricity consumption peak-valley curve set, the first-week power factor curve set, the first-week electricity consumption load curve set, and the first-week load fluctuation curve set to determine a first dispersion; if the first dispersion is less than or equal to a predetermined discrete scalar, add the first optimized metering scheme to the multiple optimized metering schemes; if the first dispersion is greater than the predetermined discrete scalar, perform a weekly electricity consumption difference analysis on multiple first users of the first user cluster; if the difference exceeds a predetermined tolerance range, set a weekly electricity consumption compensation coefficient based on the difference deviation to determine multiple weekly electricity consumption compensation coefficients; perform mapping compensation on the first optimized metering scheme based on the multiple weekly electricity consumption compensation coefficients, and then add the first optimized metering scheme to the multiple optimized metering schemes.

[0047] Specifically, a set of electricity usage characteristics for the first week of the first user cluster is obtained, including the peak and valley curves, power factor curves, load curves, and load fluctuation curves. A dispersion analysis is performed on the peak and valley curves, power factor curves, load curves, and load fluctuation curves to assess the stability of the electricity usage characteristics within the user cluster. For each electricity usage characteristic set (peak and valley curves, power factor curves, load curves, and load fluctuation curves), the average is calculated. For each data point, the square of the difference between the average value of the corresponding characteristic set is calculated. The squared differences of all data points are summed and divided by the total number of data points. The square root is then taken to obtain the overall dispersion, i.e., the first dispersion. If the dispersion is small (i.e., the user cluster has similar electricity usage behavior), the user cluster is considered to have consistent electricity usage patterns, and the first optimized metering solution can be directly applied to the user cluster.

[0048] The predetermined discrete scalar is a pre-set threshold used to assess the stability of electricity usage behavior. If the first discreteness is less than or equal to the predetermined discrete scalar, the first optimized metering scheme is added to the multiple optimized metering schemes. If the first discreteness is greater than the predetermined discrete scalar, it indicates significant electricity usage variability within the user cluster, requiring further detailed analysis. Weekly electricity usage variance analysis is performed on multiple users within the first user cluster. This involves calculating the degree of difference between each user's electricity usage characteristics and those of other users over the week. The variance is determined by calculating the variance, which reflects the average degree of difference between each user's electricity usage characteristics and those of other users. The variance of each user is compared to see if it exceeds a predetermined tolerance range. The predetermined tolerance range is a threshold used to determine whether the variance is within an acceptable range. If the variance exceeds the tolerance range, a weekly electricity compensation coefficient is set based on the variance deviation. After evaluating all variances, multiple weekly electricity compensation coefficients are determined. The weekly electricity compensation coefficient is an adjustment factor used to modify the electricity usage data of a specific user during electricity metering to more accurately reflect their electricity usage behavior.

[0049] The first optimized metering scheme is adjusted based on multiple weekly electricity consumption compensation coefficients. The compensation coefficients are mapped to the configuration parameters of the first optimized metering scheme (such as sampling frequency and accuracy). For example, users with large load fluctuations can use the compensation coefficient to increase the sampling frequency, improve metering accuracy, and more accurately capture their electricity consumption fluctuations. Users with smaller load fluctuations may reduce the sampling frequency to conserve resources. Multiple optimization schemes are generated based on the compensation coefficients to ensure that each user's actual electricity consumption characteristics are more closely aligned with the metering scheme settings. The compensated optimized metering schemes are then added to the overall optimized metering schemes. Mapping compensation adds compensation strategies to the original optimization scheme to ensure accurate and applicable metering. Through multi-level analysis and dynamic compensation adjustments, the scheme ensures that the optimized metering scheme is more adaptable to users' overall electricity consumption behavior. This helps to account for individual differences in the metering process, achieving more intelligent and accurate electricity metering results, effectively reducing the impact of electricity consumption differences on metering results, and improving the adaptability and efficiency of the metering scheme.

[0050] Furthermore, step three of this application includes:

[0051] Obtain the first month's electricity consumption information, the first week's electricity consumption information, and the first day's electricity consumption information of the first user cluster; configure a first electricity metering anomaly factor based on the first month's electricity consumption information, the first week's electricity consumption information, and the first day's electricity consumption information, and construct a first electricity metering anomaly detection branch based on the first electricity metering anomaly factor, wherein the electricity metering anomaly factor includes at least current exceeding the standard, voltage deviation, power deviation, and power factor anomaly; obtain multiple electricity metering anomaly detection branches of the multiple user clusters, and construct the electricity metering anomaly detection plug-in based on the mapping of the multiple user clusters and the multiple electricity metering anomaly detection branches.

[0052] Specifically, electricity usage information for the first user cluster at different time scales (monthly, weekly, and daily) is obtained. Each electricity usage information includes peak and valley curves, power factor curves, load curves, and load fluctuation curves. Based on the extracted electricity usage information, electricity metering anomaly factors are configured, including but not limited to current exceeding the normal operating range, voltage deviation, energy deviation, and power factor anomaly. Current exceeding the standard refers to current exceeding the normal operating range; voltage deviation refers to voltage deviation from the standard value; energy deviation refers to the difference between actual and expected energy consumption; and power factor anomaly refers to a power factor below the normal level, indicating inefficient energy utilization. Anomaly factors can be flexibly configured based on data at the monthly, weekly, and daily levels to identify subtle abnormal changes. Based on each anomaly factor, corresponding anomaly detection logic and branches are established. A first electricity metering anomaly detection branch is established in the first user cluster to detect electricity usage anomalies in that user cluster, such as excessive power fluctuations and abnormal energy usage. The above steps are repeated for each user cluster, establishing a corresponding electricity metering anomaly detection branch for each user cluster, resulting in multiple electricity metering anomaly detection branches. Each user and the corresponding electricity meter anomaly detection branch are mapped and integrated to form a unified electricity meter anomaly detection plug-in. This plug-in can flexibly adapt to the electricity usage characteristics of each user cluster, and can simultaneously receive and process electricity usage data from multiple user clusters to perform real-time anomaly detection. During operation, the electricity meter anomaly detection plug-in can continuously monitor the electricity usage of each user cluster and identify anomalies through the configuration of each branch. Once an anomaly is detected, the plug-in can trigger a corresponding alarm or automatically take corrective measures. By extracting anomaly factors from electricity usage information at different time scales and monitoring users' electricity usage behavior through detection branches, the plug-in achieves refined identification of abnormal electricity usage.

[0053] Furthermore, step 4 of this application includes:

[0054] According to the abnormality detection results, the abnormal electricity metering frequencies of the multiple user clusters are counted to determine multiple abnormal electricity metering frequencies; according to the multiple abnormal electricity metering frequencies, the multiple optimized metering schemes are dynamically adjusted, wherein, if the abnormal electricity metering frequency is less than or equal to the predetermined frequency scalar, the optimized metering scheme is targeted fine-tuned according to the abnormality detection results; if the abnormal electricity metering frequency is greater than the predetermined frequency scalar, the optimized metering scheme is overall adjusted according to the abnormality detection results.

[0055] Specifically, while the metering anomaly detection plug-in is running, abnormal metering events for each user cluster are collected and recorded in real time. The anomaly detection results for each user cluster are classified and counted, and the frequency of abnormal metering events for each user cluster is regularly counted, generating multiple abnormal metering frequency counts. Based on these counts, multiple optimized metering schemes are dynamically adjusted. Based on historical data and industry standards, a predetermined frequency scalar is pre-determined to determine the level of abnormal metering frequency. If the abnormal metering frequency is less than or equal to the predetermined frequency scalar, it indicates that the user cluster's electricity usage is relatively stable, but there may be some small-scale anomalies. Targeted fine-tuning is then implemented. While maintaining the stability of the overall solution, metering strategies for specific users are fine-tuned, such as adjusting the sampling frequency, improving data sampling accuracy, or adopting different metering methods. For example, if a user frequently experiences voltage deviations within a specific time period, a voltage compensation coefficient is set for that user to optimize their metering scheme. If the abnormal metering frequency exceeds the predetermined frequency scalar, the anomaly is considered widespread, and the optimized metering scheme for the entire user cluster needs to be re-analyzed and established. Re-evaluate the electricity usage patterns and needs of these users to identify potential systemic issues. Based on the evaluation results, establish a new optimized metering solution, including resetting metering thresholds; adjusting the electricity metering strategy, such as changing the sampling method or introducing more electricity monitoring parameters; and introducing more advanced data analysis algorithms to more accurately identify and predict abnormal behavior. After making fine-tuning or overall adjustments, continue to monitor the metering behavior of the user cluster to observe the effects of the adjustments. Regularly evaluate the effectiveness of the optimized metering solution based on implementation results and make necessary adjustments to ensure that the metering solution remains responsive to user electricity usage needs and behavior patterns. By dynamically adjusting the optimized metering solution based on statistical results of abnormal metering frequency, more efficient and accurate electricity metering management can be achieved. Targeted fine-tuning ensures that the specific needs of individual users are met, while overall adjustments help address systemic issues and maintain the effectiveness and stability of the metering solution. This not only increases the flexibility of power management but also effectively reduces the risks associated with metering anomalies, thereby improving the overall performance of the power system and user satisfaction.

[0056] In summary, the electricity metering method combined with power user grouping provided by this application has the following technical effects:

[0057] By querying the historical electricity consumption records of enterprise users in the area, users are grouped according to the historical electricity consumption records to determine multiple user clusters; daily electricity consumption intervals are divided for each of the multiple user clusters, and electricity metering strategy optimization is performed based on the division results to determine multiple optimized metering schemes; an electricity metering anomaly detection plug-in based on the multiple user clusters is established; electricity metering of enterprise users is performed according to the multiple optimized metering schemes, electricity metering anomalies are synchronously monitored through the electricity metering anomaly detection plug-in, and dynamic adjustment of the multiple optimized metering schemes is performed based on the anomaly detection results. In other words, by grouping, users are classified according to similar electricity consumption behavior patterns, and the electricity consumption data of each user cluster is further analyzed, and their daily electricity consumption is divided into different intervals. The metering strategy is customized to achieve more refined management, and abnormal electricity consumption is identified according to the historical behavior patterns of the user clusters. The metering strategy is dynamically adjusted to achieve dynamic adaptability of electricity metering, so that the metering scheme can be flexibly adjusted according to the real-time behavior of the user and the detected abnormal changes, avoiding the accumulation of long-term deviations, achieving more accurate and efficient metering of electricity consumption, and improving the pertinence and efficiency of metering.

[0058] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0059] Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. The electric meter electricity consumption measurement method combined with the electric power user grouping is characterized by: include: Query the historical electricity usage records of enterprise users in the area, group the users according to the historical electricity usage records, and determine multiple user clusters; Dividing the plurality of user clusters into daily electricity consumption intervals, performing electricity metering strategy optimization according to the division results, and determining a plurality of optimized metering schemes; Establishing an electricity meter anomaly detection plug-in based on the multiple user clusters; Executing electricity metering for enterprise users according to the multiple optimized metering schemes, synchronously monitoring electricity metering anomalies through the electricity metering anomaly detection plug-in, and dynamically adjusting the multiple optimized metering schemes according to anomaly detection results; Establishing an electricity meter anomaly detection plug-in based on the multiple user clusters, including: Obtaining electricity usage information for the first month, the first week, and the first day of a first user cluster; configuring a first electricity metering anomaly factor based on the first month's electricity usage information, the first week's electricity usage information, and the first day's electricity usage information, and establishing a first electricity metering anomaly detection branch based on the first electricity metering anomaly factor, wherein the electricity metering anomaly factor includes at least current exceeding a standard, voltage deviation, power deviation, and power factor abnormality; Acquire multiple electricity meter anomaly detection branches of the multiple user clusters, and construct the electricity meter anomaly detection plug-in based on the mapping between the multiple user clusters and the multiple electricity meter anomaly detection branches; The plurality of user clusters are respectively divided into daily electricity consumption intervals, and electricity metering strategy optimization is performed according to the division results to determine a plurality of optimized metering schemes, including: Randomly selecting a first user cluster from the multiple user clusters, and obtaining first-day electricity usage information of the first user cluster, wherein the first-day electricity usage information includes a first-day electricity usage peak-valley fitting curve, a first-day power factor fitting curve, a first-day electricity load fitting curve, and a first-day load fluctuation fitting curve; Dividing the day into multiple time intervals according to a predetermined time step length, and obtaining multiple peak-valley fitting segment centroids, multiple factor fitting segment centroids, multiple load fitting segment centroids, and multiple fluctuation fitting segment centroids of the multiple time intervals; Clustering the multiple time intervals according to the multiple peak-valley fitting segment centroids, the multiple factor fitting segment centroids, the multiple load fitting segment centroids, and the multiple fluctuation fitting segment centroids to determine multiple time thresholds and multiple power consumption feature centroid sets; Executing an electricity metering strategy optimization configuration based on the plurality of electricity feature centroid sets to determine a plurality of optimized metering strategies, wherein the electricity metering strategy includes a sampling frequency and a data sampling accuracy; constructing a first optimized metering scheme according to the multiple time thresholds and the multiple optimized metering strategies, and adding the first optimized metering scheme to the multiple optimized metering schemes; Determine multiple time thresholds, including: Arranging the multiple peak-valley fitting segment centroids in chronological order of time intervals to determine a peak-valley segment centroid sequence, and clustering the peak-valley segment centroid sequence to determine a first time threshold set; Acquire a second time threshold set, a third time threshold set, and a fourth time threshold set by clustering the multiple factor fitting segment centroids, the multiple load fitting segment centroids, and the multiple fluctuation fitting segment centroids; performing a union operation on the first time threshold set, the second time threshold set, the third time threshold set, and the fourth time threshold set in a threshold order to obtain the multiple time thresholds; The step of clustering the peak-valley segment centroid sequence to determine a first time threshold set includes: If the deviation of the centroids of two adjacent peak-valley fitting segments is less than or equal to a predetermined deviation threshold, the time intervals corresponding to the two adjacent peak-valley fitting segment centroids are merged, and the merged time interval corresponds to the average of the two adjacent peak-valley fitting segment centroids; If the deviation between the centroids of two adjacent peak-valley fitting segments is greater than a predetermined deviation threshold, a time dividing line is set and the iterative merging is continued.

2. The electric meter electricity consumption measurement method combined with electric power user grouping according to claim 1 is characterized in that: The user groups are grouped according to the historical electricity consumption records to determine multiple user clusters, including: Collecting monthly electricity consumption feature sets, weekly electricity consumption feature sets, and daily electricity consumption feature sets for multiple enterprise users based on the historical electricity consumption records, wherein the electricity consumption features include electricity peak and valley curves, power factor curves, electricity load curves, and load fluctuation curves; Extracting high-frequency features from the monthly electricity usage feature set, weekly electricity usage feature set, and daily electricity usage feature set to determine monthly electricity usage information, weekly electricity usage information, and daily electricity usage information; Using a K-means algorithm, the plurality of enterprise users are grouped according to the monthly electricity consumption information to obtain a grouping result; Performing secondary grouping on the primary grouping result according to the weekly electricity consumption information to obtain a secondary grouping result; The secondary grouping result is grouped a third time according to the daily electricity consumption information to obtain the multiple user clusters.

3. The method for metering electricity consumption in combination with power user grouping according to claim 2, characterized in that: Extracting high-frequency features from the monthly electricity consumption feature set to determine monthly electricity consumption information includes: Randomly select electricity consumption characteristics for the first month, where the electricity consumption characteristics for the first month include a peak-valley curve for electricity consumption for the first month, a power factor curve for the first month, an electricity load curve for the first month, and a load fluctuation curve for the first month; Perform similarity comparison on the first monthly electricity consumption peak-valley curve and other monthly electricity consumption peak-valley curves in the monthly electricity consumption feature set, and count the amount of data whose similarity is greater than the similarity threshold, and set it as the first centrality; If the first centrality is greater than a predetermined scalar, performing a similarity comparison between the first monthly power factor curve and other monthly power factor curves in the monthly electricity consumption feature set, and counting the amount of data with a similarity greater than a similarity threshold, and setting it as a second centrality; If the second centrality is greater than the predetermined scalar, the third centrality of the first monthly electricity load curve is calculated and obtained; if the third centrality is greater than the predetermined scalar, the fourth centrality of the first monthly load fluctuation curve is calculated and obtained; If the fourth centrality is greater than the predetermined scalar, the first monthly electricity consumption feature is added to a high-frequency monthly electricity consumption feature set, and curve fitting is performed on the high-frequency monthly electricity consumption feature set to obtain the monthly electricity consumption information.

4. The electric meter electricity consumption measurement method combined with electric power user grouping according to claim 1, characterized in that: Build the first optimized metering solution, which will also include: Obtaining a first-week electricity consumption feature set of the first user cluster, wherein the first-week electricity consumption feature set includes a first-week electricity consumption peak and valley curve set, a first-week power factor curve set, a first-week electricity consumption load curve set, and a first-week load fluctuation curve set; Performing a comprehensive dispersion calculation based on the first week's electricity peak and valley curve set, the first week's power factor curve set, the first week's electricity load curve set, and the first week's load fluctuation curve set to determine a first dispersion; If the first discreteness is less than or equal to a predetermined discrete scalar, adding the first optimized metering scheme to the plurality of optimized metering schemes; If the first discreteness is greater than a predetermined discrete scalar, performing weekly electricity usage difference analysis on the plurality of first users in the first user cluster respectively, and if the difference exceeds a predetermined tolerance interval, setting a weekly electricity usage compensation coefficient according to the difference deviation to determine a plurality of weekly electricity usage compensation coefficients; After mapping and compensating the first optimized metering scheme according to the multiple weekly electricity compensation coefficients, the first optimized metering scheme is added to the multiple optimized metering schemes.

5. The method for electricity metering combined with power user grouping according to claim 1, characterized in that: Dynamically adjusting the plurality of optimized metering schemes according to the anomaly detection results includes: counting abnormal electricity metering frequencies of the plurality of user clusters according to the abnormality detection results, and determining a plurality of abnormal electricity metering frequencies; The multiple optimized metering schemes are dynamically adjusted according to the multiple abnormal electricity metering frequencies. If the abnormal electricity metering frequency is less than or equal to the predetermined frequency scalar, the optimized metering scheme is fine-tuned in a targeted manner according to the abnormal detection result. If the abnormal electricity metering frequency is greater than the predetermined frequency scalar, the optimized metering scheme is overall adjusted according to the abnormal detection result.

Images