Metering service management method and system based on digital operation

Abstract

The application relates to the technical field of digital management of power marketing metering services, and discloses a metering service management method and system based on digital operation; the method comprises the following steps: obtaining service basic data and multi-source meter code data of metering installation and removal work orders; taking the collection time of the last period of daily frozen meter codes as a time anchor point, combining the installation and removal time and user historical power consumption data to generate an expected meter code change interval; performing standardization processing on the multi-source meter code data and generating a standardized benchmark value; determining a plurality of abnormal factors based on the standardized benchmark value, generating a metering service abnormality index; when data sources are missing, dynamically adjusting the weight according to the reliability of the remaining data sources; when the abnormality index exceeds a preset threshold value, blocking the writing of subsequent service data after the on-site input of the dismounting and returning meter codes, and generating an abnormality review task; and adjusting relevant parameters according to the review result. The application can improve the abnormality identification accuracy of the dismounting and returning meter codes and reduce the risk of abnormal meter codes entering the subsequent service link.

Description

Technical Field

[0001] This invention relates to the field of digital management technology for electricity marketing metering business, and specifically to a metering business management method and system based on digital operation. Background Technology

[0002] With the advancement of digitalization in electricity marketing, services such as meter installation and removal, meter rotation, fault meter replacement, and account cancellation are gradually being handled collaboratively through the marketing business system, metering automation system, mobile operation terminals, and metering asset management system. During the meter installation and removal process, data such as the removed meter code, photos of the electricity meter, meter code, asset code, seal, on-site location information, daily frozen meter codes, and the last meter reading code are entered on-site. This data is typically used for work order archiving, metering asset status changes, and related processing for electricity bill settlement.

[0003] In the existing metering installation and removal operations, staff typically combine on-site data entry of removed meter codes, equipment reading of meter codes, image recognition of meter codes, daily frozen meter codes, last meter reading of meter codes, and on-site evidence to manually verify or make judgments based on experience regarding whether the removed meter codes are abnormal. They then record, archive, or review the judgment results.

[0004] However, human judgment cannot simultaneously consider factors such as the time difference between the installation / removal time and the daily frozen data collection time, the user's historical electricity consumption trends, the reliability of multi-source data, and the completeness of on-site evidence. Furthermore, the judgment standards of different personnel are prone to differ. When equipment-read data, image recognition data, daily frozen data, or the last meter reading data are missing, the existing method lacks a dynamic weight adjustment mechanism based on the reliability of the remaining data sources, which can easily lead to missed or incorrect judgments. At the same time, abnormal meter codes may still enter the data links related to changes in metering asset status, installation / removal archiving, or electricity billing. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for metering business management based on digital operations, and to solve the following technical problems:

[0006] In metering installation and removal operations, given that the meter codes entered on-site may differ from those frozen daily, read from equipment, identified by image recognition, and historical electricity consumption trends, and that some data sources may be missing, how can we improve the accuracy and robustness of abnormal meter code identification and prevent abnormal meter codes from entering the asset status change, installation / removal archiving, and electricity billing data links?

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A metering business management method based on digital operations, the method comprising the following steps:

[0009] S1: Obtain the basic business data of the metering installation and dismantling work order, including the metering point number, installation and dismantling time and the electricity meter asset number;

[0010] S2: Using the electricity meter asset number and the metering point number as associated fields, obtain the multi-source meter code data corresponding to the metering installation and dismantling work order. The multi-source meter code data includes the latest daily frozen meter code and the last meter reading meter code.

[0011] S3: Using the collection time corresponding to the latest daily frozen meter code as the time anchor point, determine the time interval between the installation / removal time and the collection time, and generate the expected meter code change range corresponding to the meter installation / removal work order based on the user's historical electricity consumption data.

[0012] S4: Standardize the multi-source table code data;

[0013] S5: Based on the time interval, the user's historical electricity consumption data, the last meter reading code, and the difference between the latest daily frozen meter reading code and the last meter reading code, a standardized benchmark value is generated to evaluate the degree of meter reading anomaly; the standardized benchmark value is used to normalize the meter reading deviation for users with different electricity consumption scales.

[0014] S6: Determine multiple anomaly factors based on the standardized benchmark values;

[0015] S7: Generate a metering business anomaly index based on the multiple anomaly factors and their corresponding weights; wherein, when table code data from any source is missing, the corresponding weights are dynamically normalized and adjusted based on the credibility of the acquired data sources and the type of the missing data sources.

[0016] S8: Control the flow of metering data link of the metering installation and dismantling work order according to the metering business anomaly index; wherein, when the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data by the on-site input, and generates an anomaly review task; the subsequent business data includes metering asset status change data, installation and dismantling archive data and / or electricity bill settlement related data;

[0017] S9: Obtain the processing result of the anomaly review task, and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the processing result.

[0018] As a further aspect of the present invention: the basic business data shall at least include user ID, business type, installation / disassembly type, meter multiplier, and line information to which the metering point belongs;

[0019] The multi-source meter code data also includes at least the meter codes entered and retrieved on-site, the meter codes read by on-site equipment, and the meter codes recognized by images.

[0020] The standardization process includes standardizing table code format, converting meter multipliers, matching table code types, aligning timestamps, and marking missing data.

[0021] The abnormal factors include at least the daily freeze-up abnormal factor, the meter code growth abnormal factor, the multi-source meter code consistency abnormal factor, the historical electricity consumption trend deviation abnormal factor, and the on-site evidence integrity abnormal factor.

[0022] As a further aspect of the present invention: the standardized benchmark value is determined based on the user's historical average daily electricity consumption, the time interval between the installation / removal time and the latest daily frozen meter reading time, and the difference between the latest daily frozen meter reading and the last meter reading, and is used to normalize the meter reading deviation for users with different electricity consumption scales.

[0023] As a further aspect of the present invention: the daily frozen price inversion anomaly factor is determined based on the difference between the field-entered withdrawn table code and the most recent daily frozen table code, as well as the allowable error range; when the field-entered withdrawn table code is lower than the most recent daily frozen table code, and the difference is greater than the allowable error range, the abnormal contribution value of the daily frozen price inversion anomaly factor is determined based on the excess amount between the difference and the allowable error range.

[0024] As a further aspect of the present invention: the meter code growth anomaly factor is determined based on the meter codes recorded on-site, the meter codes frozen in the most recent period, the user's historical average daily electricity consumption, and the time interval; when the growth of the meter codes recorded on-site relative to the meter codes frozen in the most recent period is higher than the upper limit of the expected growth range determined based on historical electricity consumption data, the abnormal contribution value of the meter code growth anomaly factor is determined according to the degree to which the growth exceeds the upper limit of the expected growth range.

[0025] As a further aspect of the present invention: the multi-source code consistency anomaly factor is determined based on the difference between the code entered and retrieved on-site and the code read by the on-site equipment and the code recognized by the image, as well as the corresponding allowable deviation range;

[0026] Specifically, the difference between the on-site recorded code and the code read by the on-site equipment is determined as the first difference, and the difference between the on-site recorded code and the image recognition code is determined as the second difference.

[0027] When at least one of the first difference and the second difference exceeds the corresponding allowable deviation range, the portion exceeding the corresponding allowable deviation range is taken as the deviation exceeding the limit, and the abnormal contribution value of the multi-source table code consistency anomaly factor is determined based on the deviation exceeding the limit.

[0028] As a further aspect of the present invention: the historical electricity consumption trend deviation anomaly factor is determined based on the degree of deviation between the current cycle's average daily electricity consumption and the user's historical average daily electricity consumption; wherein, the current cycle's average daily electricity consumption is determined based on the meter code recorded on-site, the last meter reading code, and the time interval between the last meter reading time and the installation / removal time.

[0029] As a further aspect of the present invention: the on-site evidence integrity anomaly factor is determined based on the missing information between the on-site evidence to be collected and the on-site evidence actually collected; the on-site evidence includes one or more of the following: electricity meter photos, meter code photos, asset code photos, transformer photos, seal photos, meter box photos, on-site location information, and anomaly description attachments.

[0030] As a further aspect of the present invention: when any abnormal factor cannot be determined due to the lack of a data source, the abnormal factor is not directly determined as having no abnormality. Instead, the weights are reallocated based on the credibility of the data source corresponding to the remaining identifiable abnormal factors, and the metering business abnormality index is generated based on the reallocated weights.

[0031] The reliability of the data source is determined based on the data source type, acquisition method, time difference between acquisition time and installation / disassembly time, whether the data is directly read by the device, image recognition confidence level, and whether there are records of manual modification.

[0032] A metering business management system based on digital operations, the system comprising:

[0033] The work order data acquisition module is used to acquire the basic business data of metering installation and dismantling work orders. The basic business data includes at least the user number, metering point number, electricity meter asset number, business type, installation and dismantling type, installation and dismantling time, meter multiplier, and line information to which the metering point belongs.

[0034] The multi-source meter code acquisition module is used to acquire meter codes entered on-site, meter codes read by on-site equipment, meter codes recognized by image recognition, the latest daily frozen meter code from the metering automation system, and the last meter reading meter code, using the electricity meter asset number and the metering point number as associated fields.

[0035] The time anchor calibration module is used to determine the time interval between the installation / removal time and the acquisition time by taking the acquisition time corresponding to the latest daily frozen meter code as the time anchor, and to generate the expected meter code change range based on the user's historical electricity consumption data.

[0036] The data standardization module is used to unify the table code format, convert the table calculation ratio, match the table code type, align the timestamps, and mark missing data for the multi-source table code data.

[0037] An anomaly factor determination module is used to generate standardized benchmark values ​​and determine the daily freeze-up anomaly factor, meter code growth anomaly factor, multi-source meter code consistency anomaly factor, historical electricity consumption trend deviation anomaly factor, and on-site evidence integrity anomaly factor based on the standardized benchmark values.

[0038] An anomaly index generation module is used to generate a metering business anomaly index based on the anomaly factors and their corresponding weights, and to dynamically normalize and adjust the corresponding weights based on the credibility of the acquired data sources and the type of the missing data sources when table code data from any source is missing.

[0039] The data link control module is used to control the flow of metering data in the metering installation and dismantling work order according to the metering business anomaly index. When the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data and generates an anomaly review task. The subsequent business data includes metering asset status change data, installation and dismantling archive data, and / or electricity bill settlement related data.

[0040] The feedback optimization module is used to obtain the handling result of the anomaly review task and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the handling result.

[0041] The beneficial effects of this invention are:

[0042] (1) The present invention uses the collection time of the latest daily frozen meter code as the time anchor point, and combines the installation and removal time, user's historical electricity consumption data and the last meter reading to generate a standardized benchmark value. It can judge whether the change of the removed meter code is reasonable based on the actual time interval and user's electricity consumption trend, and improve the identification accuracy of daily frozen inversion anomaly, meter code growth anomaly and historical electricity consumption trend deviation anomaly.

[0043] (2) This invention standardizes the meter codes entered on-site, the meter codes read by on-site equipment, the meter codes recognized by images, the meter codes frozen in the latest daily period, and the meter codes from the last meter reading. It generates a metering business anomaly index based on the consistency of multi-source meter codes, the integrity of on-site evidence, and the credibility of the data source. When the data source is missing, it is not directly judged as having no anomaly, but rather the weights are redistributed according to the remaining identifiable anomaly factors, thereby reducing the risk of misjudgment and omission due to missing data or different data source credibility.

[0044] (3) The present invention controls the data link flow of metering installation and dismantling work orders based on the metering business anomaly index. When the anomaly index exceeds the preset threshold, it blocks the on-site entry of dismantled meter codes into metering asset status change data, installation and dismantling archive data or electricity bill settlement related data. It also adjusts the threshold parameters, weight parameters or data source credibility parameters according to the handling results of the anomaly review task, thereby reducing the possibility of anomaly codes entering the subsequent business link and improving the adaptability of subsequent anomaly identification rules. Attached Figure Description

[0045] The invention will now be further described with reference to the accompanying drawings.

[0046] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0047] Figure 2 This is a system module framework diagram of one embodiment of the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 As shown, in one embodiment, a metering business management method based on digital operations is provided, including the following steps:

[0050] S1: Obtain the basic business data of the metering installation and dismantling work order. The basic business data includes at least the user number, metering point number, electricity meter asset number, business type, installation and dismantling type, installation and dismantling time, meter multiplier, and line information to which the metering point belongs.

[0051] S2: Using the electricity meter asset number and the metering point number as associated fields, obtain multi-source meter code data corresponding to the metering installation and dismantling work order. The multi-source meter code data includes at least the meter code entered on-site, the meter code read by on-site equipment, the meter code recognized by image, the meter code frozen in the latest daily period, and the meter code read last time.

[0052] S3: Using the collection time corresponding to the latest daily frozen meter code as the time anchor point, determine the time interval between the installation / removal time and the collection time, and generate the expected meter code change range corresponding to the meter installation / removal work order based on the user's historical electricity consumption data.

[0053] S4: Standardize the multi-source table code data. The standardization process includes unifying the table code format, converting the table multiplier, matching the table code type, aligning the timestamps, and marking missing data.

[0054] S5: Based on the time interval, the user's historical electricity consumption data, the last meter reading code, and the difference between the latest daily frozen meter reading code and the last meter reading code, a standardized benchmark value is generated to evaluate the degree of meter reading anomaly; the standardized benchmark value is used to normalize the meter reading deviation for users with different electricity consumption scales.

[0055] S6: Based on the standardized benchmark value, determine multiple abnormal factors, including at least the daily freeze-up abnormal factor, meter code growth abnormal factor, multi-source meter code consistency abnormal factor, historical electricity consumption trend deviation abnormal factor, and on-site evidence integrity abnormal factor.

[0056] S7: Generate a metering business anomaly index based on the multiple anomaly factors and their corresponding weights; wherein, when table code data from any source is missing, the corresponding weights are dynamically normalized and adjusted based on the credibility of the acquired data sources and the type of the missing data sources.

[0057] S8: Control the flow of metering data link of the metering installation and dismantling work order according to the metering business anomaly index; wherein, when the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data by the on-site input, and generates an anomaly review task; the subsequent business data includes metering asset status change data, installation and dismantling archive data and / or electricity bill settlement related data;

[0058] S9: Obtain the processing result of the anomaly review task, and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the processing result;

[0059] Specifically, in step S1, the metering installation and removal work order is a work order generated during the business process of removing, replacing, rotating, replacing faulty meters, removing meters for account cancellation, or installing new meters; the basic business data can be obtained from the marketing management system, metering asset management system, or construction control system, and the following fields can be obtained:

[0060] The work order number is used to uniquely identify this metering installation and disassembly service.

[0061] User ID is used to identify the electricity customer corresponding to this service;

[0062] Username is used to display electricity customers and for manual verification;

[0063] Metering point number is used to identify the metering location corresponding to this metering service;

[0064] The electricity meter asset number is used to identify the specific electricity meter involved in this installation and removal.

[0065] The service type is used to identify whether this service is account cancellation, rotation, change, new installation, capacity expansion, or other services.

[0066] The type of installation / removal is used to identify whether this operation is an installation, removal, replacement, or a combination of installation and disassembly.

[0067] Assembly / disassembly time is used to record the actual time when assembly / disassembly occurs on site.

[0068] Meter multiplier is used to uniformly convert meter code data from different sources;

[0069] Information about the line to which the metering point belongs, used to identify the line, transformer area, or power supply range to which the metering point belongs;

[0070] User type is used to identify whether a user is a low-voltage residential user, a low-voltage non-residential user, a dedicated transformer user, a high-voltage user, or a metering point user.

[0071] The metering method is used to identify whether the metering point is measured by the main meter, sub-meter, gateway meter, or sub-metering point.

[0072] In step S2, the system uses the user ID, metering point ID, and electricity meter asset ID as joint verification fields to perform consistency verification on the work order data. When the user ID does not match the metering point ID, or the metering point ID does not match the electricity meter asset ID, the work order is marked as a basic data anomaly work order and enters the manual review process. When the basic data verification passes, the system retrieves the meter code data corresponding to the installation / removal work order from multiple data sources. The meter code data includes:

[0073] The code for the dismantled meter is entered on-site by on-site personnel using a mobile terminal.

[0074] The field equipment reads the meter reading directly from the energy meter via infrared, Bluetooth, carrier wave, RS485, or near-field communication.

[0075] The image recognition meter code is obtained by capturing a photo of the electricity meter display screen with a mobile terminal and then recognizing it using OCR.

[0076] The most recent daily frozen meter code is obtained by the metering automation system from the daily frozen data before or closest to the installation / disassembly time.

[0077] The meter code from the last meter reading, the meter code recorded in the previous meter reading cycle;

[0078] For each table code data, the system simultaneously records its data source, collection time, collection personnel or equipment number, whether there is any manual modification record, image recognition confidence level, and data reception time; if a data source fails to obtain data, the data source is marked as missing, instead of being judged as normal by default.

[0079] Specifically, the process for determining the time interval between the assembly / disassembly time and the data acquisition time is as follows:

[0080] The system uses the latest daily freeze table code R. f The corresponding acquisition time is used as the time anchor point, and this acquisition time is denoted as t. f Assembly / disassembly time t c The time interval between the daily freezing and collection time It can be obtained through the following formula:

[0081] ;

[0082] in, The unit is day; The unit is hours.

[0083] Through the above technical solution, this embodiment uses the collection time corresponding to the most recent daily frozen meter code as the time anchor point to determine the time interval between the installation / removal time and the collection time, and generates the expected meter code change range corresponding to the meter installation / removal work order based on the user's historical electricity consumption data; it can determine whether the increase in the meter code removed on site conforms to the user's normal electricity consumption pattern in that time period based on the actual time difference between the installation / removal time and the daily frozen collection time, thereby improving the accuracy of identifying abnormal meter code growth and abnormal daily frozen data;

[0084] This embodiment standardizes the multi-source table code data, thereby eliminating the incomparability issues caused by differences in format, multiplier, table code type, and acquisition time between different data sources and improving the reliability of multi-source table code verification.

[0085] This embodiment generates a standardized benchmark value for evaluating the degree of meter code anomalies based on the time interval, the user's historical electricity consumption data, the last meter reading code, and the difference between the most recent daily frozen meter reading code and the last meter reading code. This standardized benchmark value is used to normalize the meter code deviation for users with different electricity consumption scales. Based on the standardized benchmark value, it determines the daily frozen meter reading inversion anomaly factor, meter code growth anomaly factor, multi-source meter code consistency anomaly factor, historical electricity consumption trend deviation anomaly factor, and on-site evidence integrity anomaly factor. This allows for a comprehensive judgment of whether the retrieved meter code is abnormal from multiple dimensions, including short-cycle meter code changes, long-cycle historical trends, multi-source data consistency, and on-site evidence integrity. Compared to a single meter code comparison method, this reduces the risk of missed and false judgments.

[0086] In this embodiment, when meter code data from any source is missing, the corresponding weights are dynamically normalized and adjusted according to the credibility of the acquired data source and the type of the missing data source. This ensures that even in cases of incomplete data such as failure of field equipment to read data, failure of image recognition, missing daily frozen data, or missing data from the last meter reading, the remaining reliable data sources can still be used to complete the anomaly evaluation, thereby improving the adaptability and robustness of the metering installation and dismantling anomaly identification process to missing data.

[0087] This embodiment controls the flow of metering data in the metering installation and dismantling work order based on the metering business anomaly index. When the metering business anomaly index exceeds a preset threshold, the writing of the meter code to the metering asset status change data, installation and dismantling archive data, or electricity bill settlement related data by the on-site entry of the dismantled meter code is blocked. Thus, abnormal meter codes can be intercepted before entering subsequent business processes, avoiding the contamination of metering asset status, installation and dismantling archive records, and electricity bill settlement related data by erroneous dismantled meter codes, thereby improving the security and accuracy of the metering business data link.

[0088] This embodiment generates an anomaly review task when the anomaly index exceeds a preset threshold, and adjusts the threshold parameters, weight parameters, or data source credibility parameters of the anomaly factor based on the handling results of the anomaly review task. As a result, the manual review results can be fed back to the anomaly identification model, so that the system can continuously optimize the anomaly judgment rules based on the handling results of real anomalies, false anomalies, data source errors, or lack of on-site evidence, improve the accuracy of subsequent anomaly identification of similar metering installation and dismantling work orders, and reduce the probability of unnecessary business interruption.

[0089] Specifically, the handling results include one or more of the following: anomaly confirmation, false alarm confirmation, data source error, and insufficient on-site evidence;

[0090] When the handling result is confirmed as an anomaly, the threshold parameter of the corresponding anomaly factor is reduced by 5% to 15% of the current threshold, or the weight parameter is kept unchanged.

[0091] When the handling result is confirmed as a false alarm, the threshold parameter of the corresponding abnormal factor is increased by 10% to 20% of the current threshold, or the weight parameter of the corresponding abnormal factor is decreased by 5% to 15% of the current weight.

[0092] When the processing result is that the data source is incorrect, the confidence parameter of the corresponding data source is reduced by 10% to 30% of the current confidence value, and the error type of the data source is recorded.

[0093] When the outcome of the handling is that there is insufficient on-site evidence, the weight parameter of the abnormal factor of on-site evidence integrity is increased by 10% to 20% of the current weight, or the evidence weight of this evidence item in subsequent work orders is reduced.

[0094] After the parameters are adjusted, the system will save the adjustment record to the parameter update log and apply the updated parameters in the anomaly judgment of the next work order of the same type.

[0095] In summary, this embodiment can achieve closed-loop control in metering installation and removal operations, including multi-source meter code reliable verification, historical electricity consumption trend judgment, dynamic processing of missing data, abnormal index evaluation, data link interruption, and review feedback optimization. This improves the accuracy, completeness, and consistency of metering installation and removal data, and reduces the risk of abnormally removed meter codes entering the asset status change, installation and removal archiving, and electricity billing links.

[0096] As one embodiment of the present invention, the process for determining the expected range of code changes is as follows:

[0097] The system retrieves the user's historical daily electricity consumption data from the historical electricity consumption database for a preset period prior to the installation / removal time; the preset period can be the most recent 30, 60, or 90 days. After removing days of power outages, abnormal loads, meter malfunctions, and days with missing data, the system calculates the user's historical baseline daily electricity consumption Q using the following formula. basic :

[0098] ;

[0099] Among them, Q i The median is the electricity consumption on day i within a preset period; n is the total number of days in the preset period, i ∈ n; mediam is the median; Q x Q represents the median of the user's historical daily electricity consumption. y γ is the average daily electricity consumption of this user in history; γ1 is the median weighting coefficient;

[0100] As one embodiment of the present invention, the process of obtaining the median weighting coefficient γ1 is as follows:

[0101] ;

[0102] Among them, Q bd γ is the robust dispersion index; k is the dispersion correction coefficient, used to convert the dispersion quantity obtained based on the absolute deviation of the median into a scale similar to the standard deviation; the correction coefficient can also be adjusted according to the user type, metering point type, or historical electricity consumption data distribution characteristics, and k is preferably 1.4826; f(X) is the first judgment function, when X>0, f(X)=X; when X≤0, f(X)=0; min γ is the minimum value of the median weighting coefficient. min ∈[0.2,0.5];γ1∈[γ min [1]; R0 is the first preset comparison value; C1 is the first preset constant;

[0103] It should be noted that the first preset comparison value R0 is used to determine whether the fluctuation of a user's historical daily electricity consumption reaches a level that requires increasing the median weight; specifically, R0 is related to the robust dispersion index Q. bd They have the same dimensions and are both units of daily electricity consumption; when Q bd When ≤R0, it means that the user's historical daily electricity consumption fluctuation has not exceeded the preset fluctuation level, at which point f(Q) bd - R0) = 0, median weighting coefficient γ1 = γ min Historical benchmark day electricity consumption is more of a reference average; when Q bd When R0 is greater than the preset fluctuation level, it indicates that the user's historical daily electricity consumption fluctuation exceeds the preset fluctuation level, and the system adjusts accordingly with Q. bd The median weighting coefficient γ1 is increased by the increase of the value of the median; the first preset comparison value R0 can be determined according to the user type, metering point type or historical benchmark electricity consumption level.

[0104] The expected range of code changes is [Q]. basic Δt f - δ low Q basic Δt f - δ up ],in:

[0105] δ low = β low Q bd Δt f + ε low This is the downward tolerance;

[0106] Δ up = β up Q bd Δt f + ε up This is for upward tolerance;

[0107] β low and β up β is the fluctuation tolerance coefficient. low The value of β ranges from 0.5 to 1.5. up The value range is 1.5 to 3.0 (the upward tolerance is greater than the downward tolerance to cope with normal electricity consumption growth such as increased load);

[0108] It should be noted that the fluctuation tolerance coefficient β low and β up The rules for determining it are as follows:

[0109] For low-voltage residential users, β low = 0.8, β up = 2.0;

[0110] For low-voltage non-residential users, β low = 1.0, β up = 2.2;

[0111] For dedicated transformer users, β low = 1.2, β up = 2.5;

[0112] For high-voltage users or users at metering points, β low = 1.5, β up = 3.0;

[0113] When the coefficient of variation C of the user's historical daily electricity consumption v When β > 0.5, low and β up All values ​​should be adjusted by multiplying the aforementioned base values ​​by 1.2.

[0114] The upper limit of the expected growth range G max That is, the upper bound of the interval;

[0115] As one embodiment of the present invention, the process for determining the first preset comparison value R0 is as follows:

[0116] ;

[0117] Where a0 is the first fluctuation proportionality coefficient; τ q To preset the minimum daily power consumption parameter, in order to avoid when Q x If the value is too small, R0 will be too small;

[0118] It should be noted that the first fluctuation ratio coefficient a0 is configured according to the user type or metering point type; for low-voltage residential users, since their daily electricity consumption is usually relatively low and the fluctuation range is small, the value of a0 is [0.05, 0.15], preferably 0.10; for low-voltage non-residential users, since their electricity consumption behavior is greatly affected by business hours, equipment start-up and shutdown or business activities, the value of a0 is [0.1, 0.2], preferably 0.15; for dedicated transformer users, high-voltage users or metering point users, since their load scale is large and the daily electricity consumption fluctuation may be more obvious, the value of a0 is [0.2, 0.3], preferably 0.25;

[0119] Through the above technical solution, the first preset comparison value in this embodiment can adaptively change with the user's electricity consumption scale and user type, so that different user types correspond to different fluctuation judgment thresholds, avoiding setting too high a threshold for low electricity consumption users and also avoiding setting too low a threshold for high electricity consumption users.

[0120] As one embodiment of the present invention, the standardized benchmark value is determined based on the user's historical average daily electricity consumption, the time interval between the installation / removal time and the latest daily frozen meter reading time, and the difference between the latest daily frozen meter reading and the last meter reading, and is used to normalize the meter reading deviation of users with different electricity consumption scales.

[0121] Specifically, the standardized reference value E is determined in the following way:

[0122] According to the latest daily freeze code R f Meter code R from the last meter reading l The difference between them determines the electricity consumption for the reference period: Q ref = R f -R l ;

[0123] Based on the time interval Δt between the assembly / disassembly time and the daily freezing data collection time f Compared with the user's historical baseline daily electricity consumption Q basic Determine the expected increment of the table code: ΔR exp =Q basic Δt f ;

[0124] According to the expected code increment ΔR exp With the reference cycle power consumption Q ref Generate the standardized baseline value E: E = max(ΔR) exp Q ref Δt f / T ref Based on the standardized benchmark value E, the standardized deviation scale S = η × E is determined, where η is a preset proportional coefficient, and the value of η ranges from 0.1 to 0.5, used to map the benchmark value to the deviation evaluation scale; determined according to user type: for low-voltage residential users, η = 0.30 to 0.50, preferably 0.40; for low-voltage non-residential users, η = 0.20 to 0.40, preferably 0.30; for dedicated transformer users, η = 0.15 to 0.30, preferably 0.20; for high-voltage users or metering point users, η = 0.10 to 0.25, preferably 0.15;

[0125] Through the above technical solution, the standardized benchmark value in this embodiment is not determined solely by a single meter reading, but simultaneously considers the user's historical average daily electricity consumption, the time interval between the installation / removal time and the latest daily frozen meter reading time, and the difference between the latest daily frozen meter reading and the last meter reading. Therefore, the system can determine a reasonable range of meter reading changes near the installation / removal time based on the user's historical electricity consumption level and the time difference between installation and removal, avoiding misjudgments caused by applying the same absolute deviation threshold to low-consumption and high-consumption users. At the same time, the difference between the latest daily frozen meter reading and the last meter reading reflects the actual electricity consumption trend within the current meter reading cycle, allowing the standardized benchmark value to consider both short-cycle time anchors and long-cycle meter reading trends. Thus, meter reading deviations for users with different electricity consumption scales can be converted to a unified evaluation scale, improving the accuracy of judging daily frozen reading anomalies, meter reading growth anomalies, and historical electricity consumption trend deviation anomalies, and reducing false alarms and missed alarms caused by differences in user electricity consumption scale, installation / removal time, or meter reading cycle.

[0126] As one embodiment of the present invention, the daily frozen price inversion anomaly factor is determined based on the difference between the field-entered withdrawn table code and the most recent daily frozen table code and the allowable error range; when the field-entered withdrawn table code is lower than the most recent daily frozen table code and the difference is greater than the allowable error range, the abnormal contribution value of the daily frozen price inversion anomaly factor is determined based on the excess between the difference and the allowable error range.

[0127] Specifically, through formulas;

[0128] ;

[0129] Calculate the daily freeze-up anomaly factor F1;

[0130] Where, r f The most recent daily frozen meter code after conversion by metering multiplier; r m The meter readings are entered on-site after conversion by the meter multiplier; ε is the allowable error range, which is determined based on the meter accuracy, acquisition time difference, or multiplier conversion error; S is the standardized deviation scale, and S > 0. The standardized deviation scale is determined based on the standardized benchmark value generation process and is used to normalize the meter reading deviations of users with different electricity consumption scales.

[0131] When r m +ε≥r f When r indicates that the recorded table code is not lower than the latest daily frozen table code, or the difference between the two does not exceed the allowable error range, F1 = 0; m +ε<r f When the recorded code for the withdrawn form is lower than the latest daily frozen form code and the difference exceeds the allowable error range, then according to... Determine the abnormal contribution value of the daily frozen inversion anomaly factor;

[0132] Through the above technical solution, this embodiment can eliminate minor deviations caused by meter accuracy errors, differences in acquisition time, or data conversion errors, and avoid misjudging normal fluctuations within the allowable error range as inverted anomalies. At the same time, for inverted situations that exceed the allowable error range, the anomaly contribution can be quantified according to the degree of exceedance, so that the metering business anomaly index can more accurately reflect the severity of the anomaly when the retrieved meter code is lower than the daily frozen meter code, thereby improving the accuracy and stability of daily frozen inverted anomaly identification.

[0133] As one embodiment of the present invention, the meter code growth anomaly factor is determined based on the meter codes recorded on-site, the meter codes frozen in the most recent period, the user's historical average daily electricity consumption, and the time interval; when the growth of the meter codes recorded on-site relative to the meter codes frozen in the most recent period is higher than the upper limit of the expected growth range determined based on historical electricity consumption data, the abnormal contribution value of the meter code growth anomaly factor is determined according to the degree to which the growth exceeds the upper limit of the expected growth range.

[0134] Specifically, through formulas;

[0135] ;

[0136] Calculate the table code growth anomaly factor F2;

[0137] Among them, G max The upper limit of the expected growth range is determined based on the user's historical average daily electricity consumption, the time interval between the installation / removal time and the latest daily frozen meter reading collection time, the degree of historical electricity consumption fluctuation, and the allowable error.

[0138] Specifically, the upper limit of the expected growth range G max Determined in the following ways:

[0139] ;

[0140] Among them, Q basic The user's historical baseline daily electricity consumption; β is the fluctuation amplification factor, with a value ranging from 1.5 to 3.0, ε G The allowable error compensation item is determined based on the meter's accuracy class and multiplier.

[0141] The rule for determining the fluctuation amplification factor β is as follows:

[0142] Based on the coefficient of variation C of the user's historical daily electricity consumption v Sure:

[0143] When C v When ≤ 0.2, β = 1.5;

[0144] When 0.2 < C v When ≤ 0.5, β = 2.0;

[0145] When C v When the β value is greater than 0.5, β = 3.0;

[0146] Wherein, the coefficient of variation C v =Q bd / Q avg Q avg This represents the average daily electricity consumption of the user throughout history.

[0147] The allowable error compensation term ε G The rules for determining it are as follows:

[0148] ε G = 0.5×magnification + 0.1×Q basic

[0149] Wherein, the multiplier is the product of the current transformer multiplier and the voltage transformer multiplier, and is dimensionless;

[0150] When the growth of the withdrawn table code recorded on-site relative to the frozen table code of the most recent period does not exceed the upper limit of the expected growth range, the table code growth anomaly factor is zero; when the growth of the withdrawn table code recorded on-site relative to the frozen table code of the most recent period exceeds the upper limit of the expected growth range, the abnormal contribution value of the table code growth anomaly factor is determined according to the degree of exceedance.

[0151] It should be noted that the upper limit of the expected growth range can be determined based on the expected increase in electricity consumption obtained from the user's historical average daily electricity consumption and the time interval, combined with the degree of historical electricity consumption fluctuation and allowable error; the degree of historical electricity consumption fluctuation can be determined based on the user's historical daily electricity consumption data within a preset period, which can be 30, 60 or 90 days before the installation or dismantling time;

[0152] Through the above technical solution, this embodiment can determine whether the increase in the meter readings recorded on-site relative to the most recent daily frozen meter readings is within a reasonable range, based on the user's historical electricity consumption level and the time difference between the daily frozen data collection time and the installation / removal time. For cases exceeding the upper limit of the expected growth range, the abnormal contribution can be quantified according to the degree of excess, avoiding the use of only a fixed difference threshold to judge users of different electricity consumption scales, thereby improving the accuracy of meter reading anomaly identification.

[0153] As one embodiment of the present invention, the multi-source table code consistency anomaly factor is determined based on the difference between the table code entered and retrieved on site and the table code read by the on site equipment and the table code recognized by the image, as well as the corresponding allowable deviation range.

[0154] Specifically, the difference between the on-site recorded code and the code read by the on-site equipment is determined as the first difference, and the difference between the on-site recorded code and the image recognition code is determined as the second difference.

[0155] When at least one of the first difference and the second difference exceeds the corresponding allowable deviation range, the portion exceeding the corresponding allowable deviation range is taken as the deviation exceeding the limit, and the abnormal contribution value of the multi-source table code consistency anomaly factor is determined according to the deviation exceeding the limit.

[0156] Before calculation, the meter codes entered on-site, read by on-site equipment, and recognized by image should be of the same meter code type, and should be processed by the aforementioned meter multiplier conversion and timestamp alignment; data sources that cannot complete meter code type matching or timestamp alignment are marked as unavailable.

[0157] Specifically, through the formula:

[0158] ;

[0159] Calculate the first difference D between the meter reading entered on-site and the meter reading read by the on-site equipment. d The second difference D between the on-site recorded code and the image recognition code. o ;

[0160] Where, r d To read the meter readings from field equipment after conversion by meter multiplier; r o The image recognition meter code is calculated using the meter's multiplier.

[0161] Through the formula:

[0162] ;

[0163] Calculate the first deviation exceeding the limit U d Second deviation exceeding the limit U o ;

[0164] Where, σ d The first permissible deviation range is defined for the meter readings of the field equipment. This first permissible deviation range is determined based on the reading accuracy, meter multiplier, minimum meter code resolution, and equipment communication method of the field equipment; σ o The second allowable deviation range is defined based on the image recognition confidence level, the clarity of the meter photo, the shooting angle, the meter magnification, and the minimum resolution of the meter.

[0165] It should be noted that both the first and second allowable deviation ranges are determined under a unified metrological standard and are obtained by querying the preset equipment tolerance parameter table and image tolerance parameter table, respectively. The first allowable deviation range is determined based on the reading accuracy of the field equipment, the meter multiplier, the minimum resolution of the meter code, the equipment communication method, and the remaining time difference after timestamp alignment. The second allowable deviation range is determined based on the image recognition confidence level, the clarity of the meter code photo, the shooting angle, the meter multiplier, the minimum resolution of the meter code, and the remaining time difference after timestamp alignment.

[0166] As the confidence level of image recognition decreases, the second allowable deviation range is increased; when image recognition fails or the confidence level of image recognition is lower than the preset confidence threshold, the image recognition table code is marked as an unavailable data source and is not included in the abnormal contribution calculation of the second difference; the preset confidence threshold is determined based on historical image samples with correct table code annotations.

[0167] When both the field device reads the meter code and the image recognition meter code are valid, the formula is used:

[0168] ;

[0169] Calculate the multi-source table key consistency anomaly factor F3;

[0170] Wherein, ρ is the dual-source deviation enhancement coefficient, and 0 < ρ ≤ 1, which is used to improve the abnormal contribution value of the multi-source code consistency anomaly factor when the code is entered and returned on-site and deviates from the code read by the on-site equipment and the code recognized by the image; preferably, the value of ρ is in the range of 0.2 to 0.5.

[0171] When neither of the two differences exceeds the corresponding allowable deviation range, the multi-source code consistency anomaly factor is set to zero; when only one difference exceeds the corresponding allowable deviation range, the anomaly contribution value is determined according to the deviation excess amount corresponding to that difference; when both differences exceed the corresponding allowable deviation range, the basic anomaly contribution is determined by the larger deviation excess amount, and the anomaly contribution is increased according to the smaller deviation excess amount and the dual-source deviation enhancement coefficient.

[0172] When only one of the two comparison data sources is valid, the other data source is marked as missing or unavailable, and its corresponding deviation exceedance is not considered zero; the current computable anomaly contribution value of the multi-source table code consistency anomaly factor is the smaller value between the ratio of the deviation exceedance corresponding to the valid data source to the standardized deviation scale and 1; when both comparison data sources are missing or unavailable, the multi-source table code consistency anomaly factor is marked as uncomputable, and the missing type is passed to the metering business anomaly index generation step, which performs dynamic weight normalization adjustment based on the remaining identifiable anomaly factors, the corresponding data source credibility, and the missing data source type.

[0173] Through the above technical solution, this embodiment can eliminate errors in device reading, multiplier conversion, and image recognition within the allowable deviation range; for cases where the deviation exceeds the limit, the abnormal contribution is determined according to the degree of exceeding the limit, and enhancement processing is performed on cases where the deviation exceeds the limit in data from two different sources, thereby reducing the risk of misjudgment caused by errors in a single data source or missing key comparison data.

[0174] As one embodiment of the present invention, the allowable deviation range of the device reading meter code is determined according to the device reading method, user type, meter multiplier, and the time interval between the device reading time and the installation / removal time; the allowable deviation range of the image recognition meter code is determined according to the image recognition confidence level, image quality score, and the time interval between the image acquisition time and the installation / removal time.

[0175] For example, for data directly read from field equipment, the allowable deviation can be configured to be 0.5kWh to 50kWh, or 0.02% to 0.10% of the recorded and retrieved meter code; for image recognition data, the allowable deviation can be configured to be 1kWh to 100kWh, or 0.05% to 0.20% of the recorded and retrieved meter code; when the image recognition confidence decreases, the image quality score decreases, or the acquisition time interval increases, the allowable deviation is increased accordingly and the image recognition data source confidence is decreased.

[0176] As one embodiment of the present invention, the historical electricity consumption trend deviation anomaly factor is determined based on the degree of deviation between the current cycle average daily electricity consumption and the user's historical average daily electricity consumption; wherein, the current cycle average daily electricity consumption is determined based on the meter code recorded on-site, the meter code of the last meter reading, and the time interval between the last meter reading time and the installation / removal time;

[0177] Specifically, the system first acquires the field-entered meter code, the last meter reading code, the meter multiplier, the installation / removal time, and the last meter reading time. It then converts the field-entered meter code and the last meter reading code into a unified electricity consumption caliber according to the meter multiplier. The current cycle's electricity consumption is determined based on the difference between the converted field-entered meter reading and the converted last meter reading value. The current cycle's average daily electricity consumption is determined based on the time interval between the current cycle's electricity consumption and the last meter reading time and the installation / removal time. When the time interval is less than a preset minimum time interval, the preset minimum time interval is used in the calculation to avoid abnormally amplifying the current cycle's average daily electricity consumption due to an excessively short time interval.

[0178] Specifically, through the formula:

[0179] ;

[0180] Calculate the average daily electricity consumption Q for the current period. cur ;

[0181] Among them, Qtotal Δt represents the electricity consumption for the current cycle. l The time interval between the last meter reading and the installation / removal time, in days; Δt min The minimum time interval is preset.

[0182] When the meter reading entered on-site is less than the previous meter reading, the system determines whether there is a carry-over to zero. The maximum capacity of the meter is determined based on the number of electricity meters or the metering asset file. If the previous meter reading is close to the maximum capacity, the on-site entered meter reading is close to zero, and both meet the preset zero-return interval conditions, the system corrects the current cycle's electricity consumption according to the maximum capacity of the meter. Otherwise, the situation is treated as a meter reading inversion anomaly, and the historical electricity consumption trend deviation factor is marked as not being able to be directly calculated according to the normal trend.

[0183] The system acquires the user's historical daily electricity consumption data within a preset historical period prior to the installation / removal time. The preset historical period can be the most recent 30 days, 60 days, 90 days, or the most recent complete meter reading cycle. The system cleans the historical daily electricity consumption data, removing days with power outages, failed data collection days, meter malfunction days, obviously abnormal days, days with missing data, and abnormal days exceeding a preset fluctuation range. After cleaning, the system can determine the user's historical average daily electricity consumption using the average or median of the historical daily electricity consumption, and determine the historical fluctuation using the dispersion of the historical daily electricity consumption. The aforementioned average, median, and dispersion can be determined using conventional statistical methods in this field.

[0184] The system determines the historical electricity consumption trend deviation anomaly factor based on the degree of deviation between the current period's average daily electricity consumption and the user's historical average daily electricity consumption; the allowable deviation threshold is determined based on historical fluctuations and the minimum allowable average daily electricity consumption deviation; the historical electricity consumption trend deviation anomaly factor is determined according to the following formula:

[0185] ;

[0186] Calculate the historical electricity consumption trend deviation anomaly factor F4;

[0187] Among them, Q avg To represent a user's historical average daily electricity consumption; T q Indicates the allowable deviation from the threshold; S q A standardized measure of average daily electricity consumption;

[0188] Then, using the formula:

[0189] ;

[0190] Calculate the allowable deviation threshold T q Standardized Scale of Daily Electricity Consumption S q ;

[0191] Where μ is the volatility coefficient; ε min ω is the minimum permissible daily average electricity consumption deviation; ω is the proportionality coefficient.

[0192] The rule for determining the fluctuation coefficient μ is as follows:

[0193] Based on the coefficient of variation C of the user's historical daily electricity consumption v Sure:

[0194] When C v When μ ≤ 0.2, μ = 1.0;

[0195] When 0.2 < C v When μ ≤ 0.5, μ = 1.5;

[0196] When C v When μ > 0.5, μ = 2.0;

[0197] The minimum permissible daily average electricity consumption deviation ε min The rules for determining it are as follows:

[0198] For low-voltage residential users, ε min = 0.1 kWh / day;

[0199] For low-voltage non-residential users, ε min = 0.5 kWh / day;

[0200] For dedicated transformer users, ε min = 2.0 kWh / day;

[0201] For high-voltage users or users at key metering points, ε min = 5.0 kWh / day.

[0202] The rule for determining the proportionality coefficient ω is as follows:

[0203] Based on the user's historical average daily electricity consumption Q avg Sure:

[0204] When Q avg When the consumption is ≤ 1 kWh / day, ω = 0.50;

[0205] When 1 avg When the energy consumption is ≤ 10 kWh / day, ω = 0.30;

[0206] When 10 avg When the energy consumption is ≤ 100 kWh / day, ω = 0.20;

[0207] When Q avg When the daily output is >100 kWh, ω = 0.15;

[0208] ​​When the deviation between the current period's average daily electricity consumption and the user's historical average daily electricity consumption does not exceed the allowable deviation threshold, the historical electricity consumption trend deviation anomaly factor is set to 0; when the deviation exceeds the allowable deviation threshold, the system determines the anomaly contribution value based on the ratio of the excess portion to the standardized scale of the average daily electricity consumption, and the anomaly contribution value does not exceed 1.

[0209] When the last meter reading code is missing, the last meter reading time is missing, the number of historical valid electricity consumption days is less than a preset number, or the electricity consumption for the current period cannot be reliably determined, the system marks the historical electricity consumption trend deviation anomaly factor as undeterminable. When generating the metering business anomaly index, it performs dynamic weight normalization based on other determinable anomaly factors and the reliability of their data sources, and does not directly determine the anomaly factor as non-anomaly-free. The preset number can be configured according to the user type, for example, configured to be no less than 7 valid electricity consumption days.

[0210] Through the above technical solution, this embodiment can combine equipment reading data, image recognition data, on-site data entry data, and historical electricity consumption trends to perform differentiated, dynamic, and trend-based verification of the meter codes removed from the metering installation and removal work order. This improves the accuracy of abnormal meter code identification, reduces the risk of false alarms and missed alarms, and reduces the possibility of abnormal meter codes entering the subsequent asset status change, installation and removal archiving, or electricity bill settlement related data links.

[0211] As one embodiment of the present invention, the on-site evidence integrity anomaly factor is determined based on the missing information between the on-site evidence that should be collected and the on-site evidence that is actually collected; the on-site evidence includes one or more of the following: photos of electricity meters, meter codes, asset codes, transformers, seals, meter boxes, on-site location information, and anomaly description attachments;

[0212] Specifically, the system determines the list of on-site evidence to be collected for each meter installation / removal work order based on its business type, installation / removal type, user type, meter access method, and on-site operation requirements. For meter removal, rotation, fault replacement, or account closure / meter removal, the list of on-site evidence to be collected includes at least a photo of the electricity meter, a photo of the meter code, a photo of the asset code, and on-site location information. When the work order involves access via a current transformer, the list of on-site evidence to be collected also includes a photo of the current transformer. When the work order involves the removal or replacement of seals, the list of on-site evidence to be collected also includes a photo of the seals. When on-site personnel enter an anomaly description or the system identifies an anomaly in the meter code, the list of on-site evidence to be collected also includes an anomaly description attachment.

[0213] The system acquires actual on-site evidence and performs integrity verification on each piece of on-site evidence. The integrity verification includes: determining whether the evidence file exists, whether the evidence file type conforms to a preset format, whether the photo clarity meets the recognition requirements, whether the time difference between the photo acquisition time and the installation / disassembly time exceeds a preset time threshold, whether the on-site location information is within the preset location range of the metering point, whether the asset code in the photo matches the electricity meter asset number in the work order, and whether the meter code photo can identify valid meter code information.

[0214] When a piece of on-site evidence that should be collected is not collected, the document is damaged, the photo is blurry, the photo collection time is abnormal, the on-site location information is missing, the on-site location information is outside the preset range, the asset code in the photo is inconsistent with the work order, or the valid table code information cannot be identified in the table code photo, the system will mark the on-site evidence as missing or unqualified.

[0215] The system assigns corresponding weights to different types of on-site evidence; these weights are determined based on the degree to which the on-site evidence affects the authenticity of the retrieved meter codes and the on-site nature of the work order; for example, the weights of meter code photos, electricity meter photos, and asset code photos are higher than the weights of meter box photos and anomaly explanation attachments. For evidence items that are not applicable to the current work order, the system does not include them in the list of on-site evidence to be collected, nor does it participate in the calculation of the on-site evidence integrity anomaly factor.

[0216] Specifically, through the formula:

[0217] ;

[0218] Calculate the abnormality factor F5 of the integrity of on-site evidence;

[0219] Where J represents the number of on-site evidences to be collected for the current work order, j∈J; W j M represents the weight of the on-site evidence to be collected for the j-th item; j This is a marker indicating the absence or inadequacy of the on-site evidence to be collected for item j. When the on-site evidence to be collected for item j is missing or inadequate, M... j Take 1, when the j-th item is qualified to collect on-site evidence, M j Set to 0; The weighted result for missing or inadequate evidence items; This is the total weighted sum of the on-site evidence that should be collected for the current work order;

[0220] The weights can be normalized or unnormalized. When unnormalized weights are used, they are normalized by dividing by the sum of the weights of the on-site evidence to be collected in the current work order. If the current work order does not have any on-site evidence to be collected, or the sum of the weights is 0, the system does not calculate the abnormal factor and marks it as undeterminable.

[0221] When the on-site evidence integrity anomaly factor is 0, it means that all the on-site evidence required for the current work order has been collected and verified. When the on-site evidence integrity anomaly factor is greater than 0, it means that there are missing or unqualified on-site evidence in the current work order. The closer the on-site evidence integrity anomaly factor is to 1, the higher the degree of missing or unqualified on-site evidence.

[0222] In a specific example, the system identifies the meter code photo, electricity meter photo, asset code photo, seal photo, on-site location information, and anomaly description attachment as the on-site evidence to be collected for the current work order, and assigns weights to each. If the meter code photo is missing, the on-site location information is outside the preset range, and all other evidence items are qualified, the system calculates the on-site evidence integrity anomaly factor based on the weights corresponding to the meter code photo and on-site location information, and uses this factor as the input parameter for generating the metering business anomaly index.

[0223] When key on-site evidence is missing, the system can further increase the abnormal risk level of the work order. The key on-site evidence includes one or more of the following: meter reading photos, electricity meter photos, asset code photos, and on-site location information. If key on-site evidence is missing or unqualified, the system can set the on-site evidence integrity anomaly factor to no less than a preset lower limit, or generate an on-site evidence supplementation task.

[0224] Specifically, the weight of the on-site evidence can be configured according to the following rules:

[0225]

[0226] The above weights are unnormalized weights, and the sum of all weights is 100. When normalized weights are used, each of the above weights is divided by 100.

[0227] When a certain evidence item is not in the list of items to be collected due to the inapplicability of the work order type, that evidence item will not participate in the calculation of the total weight.

[0228] The key on-site evidence includes photos of meter readings (W1=25), photos of electricity meters (W2=20), photos of asset codes (W1=20), and on-site location information (W4=15); if any of the key on-site evidence is missing or unqualified, the on-site evidence integrity anomaly factor F_5 shall be no less than 0.5.

[0229] Through the above technical solution, this embodiment can transform situations such as missing on-site evidence, unqualified evidence quality, mismatch between evidence and work order, and abnormal on-site positioning into quantifiable abnormal factors, so that the integrity of on-site evidence can participate in the abnormal identification of metering and dismantling work orders, reducing the risk of misjudgment caused by relying solely on the on-site input of dismantled meter codes for judgment.

[0230] As one embodiment of the present invention, when any abnormal factor cannot be determined due to the lack of data source, the abnormal factor is not directly determined as having no abnormality. Instead, the weights are reallocated based on the credibility of the data source corresponding to the remaining identifiable abnormal factors, and the metering business abnormality index is generated based on the reallocated weights.

[0231] The reliability of the data source is determined based on the data source type, acquisition method, time difference between acquisition time and installation / disassembly time, whether the data is directly read by the device, image recognition confidence level, and whether there are records of manual modification.

[0232] Specifically, the system pre-configures initial weights for each abnormal factor; the abnormal factors include daily frozen inversion abnormal factor, meter code growth abnormal factor, multi-source meter code consistency abnormal factor, historical electricity consumption trend deviation abnormal factor, and on-site evidence integrity abnormal factor; before calculating the metering business abnormal index, the system determines whether the data required for each abnormal factor is complete; when the meter code data, historical electricity consumption data, image recognition data, or on-site evidence data required for a certain abnormal factor are missing, making it impossible to reliably calculate the abnormal factor, the abnormal factor is marked as uncertain;

[0233] Specifically, through the formula:

[0234] ;

[0235] Calculate the credibility C of the data source s ;

[0236] Among them, C min To preset the minimum confidence level; C base The base credibility corresponding to the data source type; R t R is the time correction factor; d R is the direct-read correction factor for the equipment. o R is the image recognition correction coefficient; m The correction factor was manually adjusted.

[0237] The basic credibility is determined based on the data source type; the basic credibility of data directly read from field equipment and daily frozen data from the metering automation system is higher than that of data entered on-site; the basic credibility of image recognition data is determined based on the image recognition results; the basic credibility of on-site evidence data is determined based on the evidence type and evidence quality. The time correction coefficient is determined based on the time difference between the data acquisition time and the installation / disassembly time; the larger the time difference, the smaller the time correction coefficient. The equipment direct reading correction coefficient indicates whether the data was directly read by the equipment; when the data is directly read by the field equipment and has not been manually modified, the equipment direct reading correction coefficient is higher. The image recognition correction coefficient is determined based on the image recognition confidence level and image quality score; for non-image recognition data, the image recognition correction coefficient is 1. The manual modification correction coefficient is determined based on whether there are records of manual modification; when there are records of manual modification, the manual modification correction coefficient is reduced. The values ​​of the above correction coefficients range from 0 to 1.

[0238] For each identifiable anomaly factor, the system determines the corresponding anomaly factor credibility based on the credibility of the data source it relies on. For example, the daily frozen data inversion anomaly factor corresponds to on-site data entry and daily frozen data; the multi-source meter code consistency anomaly factor corresponds to on-site data entry, on-site equipment reading data, and image recognition data; the historical electricity consumption trend deviation anomaly factor corresponds to on-site data entry, the last meter reading data, and historical electricity consumption data; the on-site evidence integrity anomaly factor corresponds to on-site evidence data. The credibility of anomaly factors can be the minimum value among the credibility of the data sources they rely on, or it can be determined by weighting according to a preset ratio.

[0239] Let A be the set of identifiable anomalies, and then use the formula:

[0240] ;

[0241] Calculate the reassigned weight W of the k-th determinable outlier. ks ;

[0242] Where K is the number of identifiable outliers in set A, k∈K; W k0 C represents the initial weight of the k-th determinable anomaly factor; k The confidence level of the k-th identifiable anomaly factor; It is the sum of the products of the initial weights and confidence levels of each identifiable anomaly factor in set A;

[0243] Based on the redistributed weights, using the formula:

[0244] ;

[0245] Calculate the metering business anomaly index I;

[0246] Among them, Fk This represents the value of the k-th determinable anomaly factor. When an anomaly factor is undeterminable, it does not participate in the aforementioned weight reallocation and the calculation of the metering business anomaly index. For anomalies that are undeterminable due to the lack of key data sources, the system can simultaneously generate a missing data marker and use this marker as the trigger condition for the anomaly review task. The key data sources include one or more of the following: field equipment read data, image recognition data, daily frozen data, last meter reading data, and meter code photos.

[0247] When all abnormal factors are uncertain, or when the sum of the products of the initial weights and confidence levels of each determinable abnormal factor in set A is 0, the system does not generate a normal conclusion, but instead marks the metering assembly and disassembly work order as a data-insufficient work order and generates an abnormal review task.

[0248] Through the above technical solution, this embodiment can avoid directly judging the corresponding abnormal factor as having no abnormality due to the lack of a certain data source, and can make the data source with higher credibility occupy a higher weight in the calculation of the abnormality index, thereby improving the reliability of the metering business abnormality index in the scenario of missing data.

[0249] Please see Figure 2 As shown, a metering business management system based on digital operations includes:

[0250] The work order data acquisition module is used to acquire the basic business data of metering installation and dismantling work orders. The basic business data includes at least the user number, metering point number, electricity meter asset number, business type, installation and dismantling type, installation and dismantling time, meter multiplier, and line information to which the metering point belongs.

[0251] The multi-source meter code acquisition module is used to acquire meter codes entered on-site, meter codes read by on-site equipment, meter codes recognized by image recognition, the latest daily frozen meter code from the metering automation system, and the last meter reading meter code, using the electricity meter asset number and the metering point number as associated fields.

[0252] The time anchor calibration module is used to determine the time interval between the installation / removal time and the acquisition time by taking the acquisition time corresponding to the latest daily frozen meter code as the time anchor, and to generate the expected meter code change range based on the user's historical electricity consumption data.

[0253] The data standardization module is used to unify the table code format, convert the table calculation ratio, match the table code type, align the timestamps, and mark missing data for the multi-source table code data.

[0254] An anomaly factor determination module is used to generate standardized benchmark values ​​and determine the daily freeze-up anomaly factor, meter code growth anomaly factor, multi-source meter code consistency anomaly factor, historical electricity consumption trend deviation anomaly factor, and on-site evidence integrity anomaly factor based on the standardized benchmark values.

[0255] An anomaly index generation module is used to generate a metering business anomaly index based on the anomaly factors and their corresponding weights, and to dynamically normalize and adjust the corresponding weights based on the credibility of the acquired data sources and the type of the missing data sources when table code data from any source is missing.

[0256] The data link control module is used to control the flow of metering data in the metering installation and dismantling work order according to the metering business anomaly index. When the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data and generates an anomaly review task. The subsequent business data includes metering asset status change data, installation and dismantling archive data, and / or electricity bill settlement related data.

[0257] The feedback optimization module is used to obtain the handling result of the anomaly review task and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the handling result.

[0258] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A metering business management method based on digital operations, characterized in that, The method includes the following steps: S1: Obtain the basic business data of the metering installation and dismantling work order, including the metering point number, installation and dismantling time and the electricity meter asset number; S2: Using the electricity meter asset number and the metering point number as associated fields, obtain the multi-source meter code data corresponding to the metering installation and dismantling work order. The multi-source meter code data includes the latest daily frozen meter code and the last meter reading meter code. S3: Using the collection time corresponding to the latest daily frozen meter code as the time anchor point, determine the time interval between the installation / removal time and the collection time, and generate the expected meter code change range corresponding to the meter installation / removal work order based on the user's historical electricity consumption data. S4: Standardize the multi-source table code data; S5: Based on the time interval, the user's historical electricity consumption data, the last meter reading code, and the difference between the latest daily frozen meter reading code and the last meter reading code, a standardized benchmark value is generated to evaluate the degree of meter reading anomaly; the standardized benchmark value is used to normalize the meter reading deviation for users with different electricity consumption scales. S6: Determine multiple anomaly factors based on the standardized benchmark values; S7: Generate a metering business anomaly index based on the multiple anomaly factors and their corresponding weights; wherein, when table code data from any source is missing, the corresponding weights are dynamically normalized and adjusted based on the credibility of the acquired data sources and the type of the missing data sources. S8: Control the flow of metering data link of the metering installation and dismantling work order according to the metering business anomaly index; wherein, when the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data and generates an anomaly review task. S9: Obtain the processing result of the anomaly review task, and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the processing result.

2. The metering business management method based on digital operation according to claim 1, characterized in that, The basic business data also includes at least the user ID, business type, installation / removal type, meter multiplier, and information about the line to which the metering point belongs; The multi-source meter code data also includes at least the meter codes entered and retrieved on-site, the meter codes read by on-site equipment, and the meter codes recognized by images. The standardization process includes standardizing table code format, converting meter multipliers, matching table code types, aligning timestamps, and marking missing data. The abnormal factors include at least the daily freeze-up abnormal factor, the meter code growth abnormal factor, the multi-source meter code consistency abnormal factor, the historical electricity consumption trend deviation abnormal factor, and the on-site evidence integrity abnormal factor.

3. The metering business management method based on digital operation according to claim 2, characterized in that, The standardized benchmark value is determined based on the user's historical average daily electricity consumption, the time interval between the installation / removal time and the latest daily frozen meter reading time, and the difference between the latest daily frozen meter reading and the last meter reading. It is used to normalize and evaluate the meter reading deviation for users with different electricity consumption scales.

4. The metering business management method based on digital operation according to claim 3, characterized in that, The daily frozen price inversion anomaly factor is determined based on the difference between the field-entered withdrawn table code and the most recent daily frozen table code, as well as the allowable error range. When the field-entered withdrawn table code is lower than the most recent daily frozen table code, and the difference is greater than the allowable error range, the abnormal contribution value of the daily frozen price inversion anomaly factor is determined based on the excess amount between the difference and the allowable error range.

5. The metering business management method based on digital operation according to claim 4, characterized in that, The meter code growth anomaly factor is determined based on the meter codes recorded on-site, the meter codes frozen in the most recent period, the user's historical average daily electricity consumption, and the time interval. When the growth of the meter codes recorded on-site relative to the meter codes frozen in the most recent period exceeds the upper limit of the expected growth range determined based on historical electricity consumption data, the abnormal contribution value of the meter code growth anomaly factor is determined according to the degree to which the growth exceeds the upper limit of the expected growth range.

6. The metering business management method based on digital operation according to claim 5, characterized in that, The multi-source code consistency anomaly factor is determined based on the difference between the code entered and retrieved on-site and the code read by the on-site equipment and the code recognized by the image, as well as the corresponding allowable deviation range. Specifically, the difference between the on-site recorded code and the code read by the on-site equipment is determined as the first difference, and the difference between the on-site recorded code and the image recognition code is determined as the second difference. When at least one of the first difference and the second difference exceeds the corresponding allowable deviation range, the portion exceeding the corresponding allowable deviation range is taken as the deviation exceeding the limit, and the abnormal contribution value of the multi-source table code consistency anomaly factor is determined based on the deviation exceeding the limit.

7. The metering business management method based on digital operation according to claim 6, characterized in that, The historical electricity consumption trend deviation anomaly factor is determined based on the degree of deviation between the current cycle's average daily electricity consumption and the user's historical average daily electricity consumption; wherein, the current cycle's average daily electricity consumption is determined based on the meter reading recorded on-site, the last meter reading, and the time interval between the last meter reading and the installation / removal time.

8. The metering business management method based on digital operation according to claim 7, characterized in that, The on-site evidence integrity anomaly factor is determined based on the missing information between the on-site evidence that should be collected and the on-site evidence that is actually collected; the on-site evidence includes one or more of the following: photos of electricity meters, meter codes, asset codes, transformers, seals, meter boxes, on-site location information, and anomaly description attachments.

9. The metering business management method based on digital operation according to claim 8, characterized in that, When any abnormal factor cannot be determined due to the lack of data source, the abnormal factor is not directly determined to be without abnormality. Instead, the weights are reallocated according to the credibility of the data source corresponding to the remaining identifiable abnormal factors, and the metering business abnormality index is generated based on the reallocated weights. The reliability of the data source is determined based on the data source type, acquisition method, time difference between acquisition time and installation / disassembly time, whether the data is directly read by the device, image recognition confidence level, and whether there are records of manual modification.

10. A metering business management system based on digital operation, applicable to the metering business management method based on digital operation as described in any one of claims 1-9, characterized in that, The system includes: The work order data acquisition module is used to acquire the basic business data of metering installation and dismantling work orders. The basic business data includes at least the user number, metering point number, electricity meter asset number, business type, installation and dismantling type, installation and dismantling time, meter multiplier, and line information to which the metering point belongs. The multi-source meter code acquisition module is used to acquire meter codes entered on-site, meter codes read by on-site equipment, meter codes recognized by image recognition, the latest daily frozen meter code from the metering automation system, and the last meter reading meter code, using the electricity meter asset number and the metering point number as associated fields. The time anchor calibration module is used to determine the time interval between the installation / removal time and the acquisition time by taking the acquisition time corresponding to the latest daily frozen meter code as the time anchor, and to generate the expected meter code change range based on the user's historical electricity consumption data. The data standardization module is used to unify the table code format, convert the table calculation ratio, match the table code type, align the timestamps, and mark missing data for the multi-source table code data. An anomaly factor determination module is used to generate standardized benchmark values ​​and determine the daily freeze-up anomaly factor, meter code growth anomaly factor, multi-source meter code consistency anomaly factor, historical electricity consumption trend deviation anomaly factor, and on-site evidence integrity anomaly factor based on the standardized benchmark values. An anomaly index generation module is used to generate a metering business anomaly index based on the anomaly factors and their corresponding weights, and to dynamically normalize and adjust the corresponding weights based on the credibility of the acquired data sources and the type of the missing data sources when table code data from any source is missing. The data link control module is used to control the flow of metering data in the metering installation and dismantling work order according to the metering business anomaly index. When the metering business anomaly index exceeds a preset threshold, the data link control module blocks the writing of the dismantled meter code into subsequent business data and generates an anomaly review task. The subsequent business data includes metering asset status change data, installation and dismantling archive data, and / or electricity bill settlement related data. The feedback optimization module is used to obtain the handling result of the anomaly review task and adjust the threshold parameter, weight parameter or data source credibility parameter of the anomaly factor according to the handling result.

Images