CN-115293257-B - Detection method and system for abnormal electricity utilization user

CN115293257BCN 115293257 BCN115293257 BCN 115293257BCN-115293257-B

Abstract

The invention provides a detection method and a detection system for abnormal electricity users, which belong to the technical field of abnormal electricity detection, and the method combines big data and a neural network technology to be applied to the detection of abnormal electricity, trains a neural network model by utilizing a large amount of historical electricity data and corresponding abnormal rates, realizes the accurate prediction of the abnormal rate of each electricity user through the trained neural network model, and can accurately locate the users with abnormal electricity; and the abnormal electricity consumption analysis is carried out according to the electricity consumption data of the abnormal electricity consumption user, so that the abnormal electricity consumption description of the abnormal electricity consumption user is clearly known, and whether other high-tech electricity stealing methods such as undervoltage electricity stealing, undercurrent electricity stealing and phase-shifting electricity stealing exist in the abnormal electricity consumption user can be detected.

Inventors

LI JIAN
Ren Xibi
CHEN YUHUI
WANG BAOLU
HUANG QI
HU WEIHAO
ZHANG ZHENYUAN
YI JIANBO
JING SHI
LI KE

Assignees

电子科技大学

Dates

Publication Date: 20260512
Application Date: 20220802

Claims (8)

1. A detection method for a user using electricity abnormally, the detection method comprising: The method comprises the steps of constructing a training data set, wherein the training data set comprises a plurality of pieces of historical electricity utilization data and an abnormality rate corresponding to the historical electricity utilization data, and the abnormality rate represents the probability that the historical electricity utilization data is abnormal electricity utilization data; Taking the historical electricity consumption data as the input of the abnormal rate prediction model, taking the abnormal rate corresponding to the historical electricity consumption data as the target output of the abnormal rate prediction model, and training the abnormal rate prediction model to obtain a trained abnormal rate prediction model; Collecting electricity consumption data of a target user in any time period, and inputting the electricity consumption data into the trained abnormal rate prediction model to obtain the abnormal rate of the target user in the time period; if the abnormal rate of the target user in the time period is higher than an abnormal threshold value, performing abnormal electricity analysis according to the electricity consumption data of the target user to obtain abnormal electricity consumption description of the target user, wherein the abnormal electricity consumption description comprises at least one of undervoltage electricity stealing, undercurrent electricity stealing, phase shifting electricity stealing or other electricity stealing methods; constructing a training data set, specifically comprising: Collecting a plurality of pieces of conventional electricity consumption data, wherein the conventional electricity consumption data comprise user side current, user side voltage and user side power in any time period; respectively carrying out normalization processing on each conventional power consumption data and calculating unbalance rate data corresponding to the conventional power consumption data, wherein the unbalance rate data comprises a current unbalance rate, a voltage unbalance rate and a power unbalance rate, and the power unbalance rate comprises an A-phase power unbalance rate, a B-phase power unbalance rate and a C-phase power unbalance rate; Respectively combining each piece of conventional electricity consumption data and unbalance rate data corresponding to the conventional electricity consumption data into one piece of historical electricity consumption data; Marking the corresponding abnormal rate of each historical electricity consumption data according to a preset rule; And after acquiring electricity consumption data of the target user in any time period and inputting the electricity consumption data into the trained abnormal rate prediction model to obtain the abnormal rate of the target user in the time period, the detection method further comprises the following steps: Acquiring an electricity utilization trend curve of a gateway where the target user is located; And if the abnormal rate of the target user in the time period after correction is higher than an abnormal threshold value, carrying out abnormal electricity analysis according to the electricity data of the target user to obtain abnormal electricity description of the target user.
2. The detection method according to claim 1, wherein, after collecting electricity data of the target user in any period of time and inputting the electricity data into the trained abnormality rate prediction model to obtain the abnormality rate of the target user in the period of time, the detection method further comprises: Predicting the abnormal rate of the target user in each time period of the day through the trained abnormal rate prediction model; According to the anomaly rates of the target user in each time period of the day, calculating the comprehensive anomaly rate of the target user in the same day, wherein the comprehensive anomaly rate is the average value of the anomaly rates of the target user in each time period of the day; And determining the current-day duration anomaly time of the target user according to the anomaly rate of the target user in each time period of the current day, wherein the current-day duration anomaly time is the sum of a plurality of time periods with the anomaly rate higher than an anomaly threshold value.
3. The detection method according to claim 2, wherein after predicting the abnormality rate of the target user in each time period of the day by the trained abnormality rate prediction model, the detection method further comprises: Acquiring historical electricity consumption of the target user in a target time period of seven days in the past, the same day of the last month and the same day of the last year, wherein the target time period is any time period with an abnormality rate higher than an abnormality threshold value; according to the historical electricity consumption of the target user in the target time period, predicting to obtain the predicted electricity consumption of the target user in the target time period; and obtaining the predicted electricity larceny amount of the target user in the target time period according to the predicted electricity consumption of the target user in the target time period and the electricity actual measurement of the target user in the target time period, wherein the electricity actual measurement is the electricity consumption of the target user in the target time period measured by an electricity meter.
4. The method of claim 1, wherein the imbalance rate data corresponding to the regular electricity usage data is calculated according to the following equation: Wherein i_xnx is the current imbalance ratio, u_xnx is the voltage imbalance ratio, p_bph is the power imbalance ratio, a_bph is the a phase power imbalance ratio, b_bph is the B phase power imbalance ratio, c_bph is the C phase power imbalance ratio, U lmax is the maximum voltage, U lmin is the minimum voltage, I lmax is the maximum current, I lmin is the minimum current, PA is the a phase power, PB is the B phase power, PC is the C phase power, PZ is the three phase power average, UA is the a phase voltage, IA is the a phase current, UB is the B phase voltage, IB is the B phase current, UC is the C phase voltage, IC is the C phase current.
5. The method of claim 1, wherein the conventional electricity usage data is normalized according to the following equation: Wherein x' is the data subjected to normalization processing, x is the conventional electricity consumption data, x min is the minimum value of the conventional electricity consumption data, and x max is the maximum value of the conventional electricity consumption data.
6. The detection method according to claim 1, wherein the determining the anomaly rate corresponding to each of the historical electricity consumption data markers according to the preset rule specifically includes: Setting initial abnormal rates for the historical electricity consumption data according to the user information corresponding to the historical electricity consumption data; and carrying out abnormal electricity utilization analysis according to the user side current, the user side voltage, the user side power, the current unbalance rate, the voltage unbalance rate and the power unbalance rate, and updating the abnormal rate of the historical electricity utilization data.
7. The method according to claim 6, wherein the setting an initial abnormality rate for each of the historical electricity consumption data according to the user information corresponding to each of the historical electricity consumption data, specifically comprises: judging whether a user corresponding to any historical electricity consumption data has abnormal electricity consumption in the past according to any historical electricity consumption data to obtain a first judgment result; If the first judgment result is negative, setting the initial abnormality rate of the historical electricity consumption data to be 0; If the first judgment result is yes, setting the initial abnormality rate of the historical electricity consumption data to be a numerical value between 0 and 1.
8. A detection system for abnormal electricity users, for implementing the detection method according to any one of claims 1to 7, comprising: the system comprises a data set construction module, a data set generation module and a data set generation module, wherein the data set construction module is used for constructing a training data set, the training data set comprises a plurality of pieces of historical electricity utilization data and an abnormality rate corresponding to the historical electricity utilization data, and the abnormality rate represents the probability that the historical electricity utilization data is abnormal electricity utilization data; the model training module is used for training the abnormal rate prediction model by utilizing the training data set to obtain a trained abnormal rate prediction model, wherein the historical electricity consumption data is used as the input of the abnormal rate prediction model, and the abnormal rate corresponding to the historical electricity consumption data is used as the target output of the abnormal rate prediction model; The abnormal rate prediction module is used for collecting electricity consumption data of a target user in any time period, inputting the electricity consumption data into the trained abnormal rate prediction model, and obtaining the abnormal rate of the target user in the time period; And the abnormal electricity utilization description module is used for carrying out abnormal electricity utilization analysis according to the electricity utilization data of the target user when the abnormal rate of the target user in the time period is higher than an abnormal threshold value to obtain abnormal electricity utilization description of the target user, wherein the abnormal electricity utilization description comprises at least one of undervoltage electricity utilization, undercurrent electricity utilization, phase-shifting electricity utilization or other electricity utilization methods.

Description

Detection method and system for abnormal electricity utilization user Technical Field The invention relates to the technical field of abnormal electricity utilization detection, in particular to a detection method and system for abnormal electricity utilization users. Background Along with diversified collection of electricity utilization user data and deepening application of the data, a database for storing the electricity utilization user data is constructed, meanwhile, the electricity utilization condition of the user can be obtained more accurately through data analysis of the electricity utilization user, at present, the traditional method for determining abnormal electricity utilization users is time-consuming and labor-consuming, all users are difficult to cover, and the current line loss condition of the whole area is difficult to accurately locate through a line loss statistical analysis function of an area of a metering automation system, but the novel electricity stealing method lacks strict logic judgment and accurate data analysis. The Chinese patent application publication No. CN113452145A discloses a method and a system for detecting the electricity consumption condition of users in a low-voltage area, which are used for acquiring the electricity consumption attribute of all users in a target area, and carrying out electricity larceny judgment on the users by combining power consumption, uncapping time, abnormal wiring and historical metering data. Disclosure of Invention The invention aims to provide a detection method and a detection system for abnormal electricity utilization users, which improve the accuracy of abnormal electricity utilization detection. In order to achieve the above object, the present invention provides the following solutions: a detection method for abnormal electricity users comprises the following steps: The method comprises the steps of constructing a training data set, wherein the training data set comprises a plurality of pieces of historical electricity utilization data and an abnormality rate corresponding to the historical electricity utilization data, and the abnormality rate represents the probability that the historical electricity utilization data is abnormal electricity utilization data; Taking the historical electricity consumption data as the input of the abnormal rate prediction model, taking the abnormal rate corresponding to the historical electricity consumption data as the target output of the abnormal rate prediction model, and training the abnormal rate prediction model to obtain a trained abnormal rate prediction model; Collecting electricity consumption data of a target user in any time period, and inputting the electricity consumption data into the trained abnormal rate prediction model to obtain the abnormal rate of the target user in the time period; If the abnormal rate of the target user in the time period is higher than an abnormal threshold value, abnormal electricity utilization analysis is carried out according to the electricity utilization data of the target user to obtain abnormal electricity utilization description of the target user, wherein the abnormal electricity utilization description comprises at least one of undervoltage electricity stealing, undercurrent electricity stealing, phase shifting electricity stealing or other electricity stealing methods. Optionally, the electricity consumption data of the target user in any time period is collected and input into the trained anomaly rate prediction model, and after the anomaly rate of the target user in the time period is obtained, the detection method further includes: Predicting the abnormal rate of the target user in each time period of the day through the trained abnormal rate prediction model; According to the anomaly rates of the target user in each time period of the day, calculating the comprehensive anomaly rate of the target user in the same day, wherein the comprehensive anomaly rate is the average value of the anomaly rates of the target user in each time period of the day; And determining the current-day duration anomaly time of the target user according to the anomaly rate of the target user in each time period of the current day, wherein the current-day duration anomaly time is the sum of a plurality of time periods with the anomaly rate higher than an anomaly threshold value. Optionally, after the trained abnormality rate prediction model predicts the abnormality rate of the target user in each time period of the day, the detection method further includes: Acquiring historical electricity consumption of the target user in a target time period of seven days in the past, the same day of the last month and the same day of the last year, wherein the target time period is any time period with an abnormality rate higher than an abnormality threshold value; according to the historical electricity consumption of the target user in the target time period, predicting to o