CN113514695B - Detection system and detection method suitable for group fixed proportion electricity stealing behavior - Google Patents

Detection system and detection method suitable for group fixed proportion electricity stealing behavior Download PDF

Info

Publication number
CN113514695B
CN113514695B CN202110237945.5A CN202110237945A CN113514695B CN 113514695 B CN113514695 B CN 113514695B CN 202110237945 A CN202110237945 A CN 202110237945A CN 113514695 B CN113514695 B CN 113514695B
Authority
CN
China
Prior art keywords
electricity stealing
user
electricity
users
detection module
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110237945.5A
Other languages
Chinese (zh)
Other versions
CN113514695A (en
Inventor
彭彦林
吴亚龙
许越杰
康锦萍
赵海森
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
North China Electric Power University
Original Assignee
North China Electric Power University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by North China Electric Power University filed Critical North China Electric Power University
Priority to CN202110237945.5A priority Critical patent/CN113514695B/en
Publication of CN113514695A publication Critical patent/CN113514695A/en
Application granted granted Critical
Publication of CN113514695B publication Critical patent/CN113514695B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R22/00—Arrangements for measuring time integral of electric power or current, e.g. electricity meters
    • G01R22/06—Arrangements for measuring time integral of electric power or current, e.g. electricity meters by electronic methods
    • G01R22/061—Details of electronic electricity meters
    • G01R22/066—Arrangements for avoiding or indicating fraudulent use
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R11/00—Electromechanical arrangements for measuring time integral of electric power or current, e.g. of consumption
    • G01R11/02—Constructional details
    • G01R11/24—Arrangements for avoiding or indicating fraudulent use

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of power utilization behavior detection and analysis of users of a power system, and relates to a detection system and a detection method suitable for group fixed proportion electricity stealing behavior. The method comprises the following steps: and converting the population fixed proportion electricity stealing detection into an optimization problem, and solving the problem based on the incremental property of the electricity consumption of the fixed proportion electricity stealing users and the existence of the line loss covariance of the transformer area. The detection system comprises: the device comprises a preprocessing module, a detection module and a judgment module. The preprocessing module extracts the electric quantity of line loss of users and distribution areas in unit time, and vectorizes and standardizes the electric quantity; the detection module outputs the electricity stealing suspicion user set by the method; the judgment module calculates the suspicion of electricity stealing of the user. The method overcomes the defect of low recall ratio in a group fixed proportion electricity stealing scene in the prior art, and can comprehensively and accurately detect the target user. Considering that most of electricity stealing of physical attack is electricity stealing with fixed proportion, and the current electricity stealing is in the trend of grouping and unitizing, the invention can effectively guide the work of anti-electricity stealing of a power grid company.

Description

Detection system and detection method suitable for group fixed proportion electricity stealing behavior
Technical Field
The invention belongs to the technical field of power utilization behavior detection and analysis of users of a power system, and relates to a detection system and a detection method suitable for group fixed proportion electricity stealing behavior.
Background
The electricity stealing behavior of the user will seriously damage the economic benefits of the power supply company, and meanwhile, the potential safety hazards of power failure, equipment damage and casualties are buried. In recent years, users have increasingly shown a tendency of group and group integration when stealing electricity, and the situation of anti-electricity-stealing facing power grid companies is more severe. As an important component of the smart grid, a high-level measuring system (such as a sensor, a smart meter and the like) provides massive power utilization data of users, so that technologies such as data mining and the like have great use in the field of electricity stealing detection. The existing data-supported electricity stealing detection methods can be divided into three categories: game theory based, electricity usage behavior pattern based and system state based. The method based on the game theory does not have the condition of engineering practicability up to now; the method based on the electricity usage behavior pattern is not very good at detecting users who steal electricity in a fixed proportion because the electricity curve of the meter of the users who steal electricity in a fixed proportion is not much different from the electricity curve of normal electricity after the standardized processing; the line loss correlation analysis method based on the system state is good at the detection of small-scale fixed proportion electricity stealing users, but when group fixed proportion electricity stealing occurs in an analyzed station area, the recall ratio is generally low.
Considering that most electricity stealing modes for tampering the electricity meter from hardware are represented by electricity stealing with a fixed proportion, and the current electricity stealing mode increasingly shows the trend of grouping and unitizing. Therefore, aiming at the electricity stealing scene, the detection system and the detection method suitable for the group fixed proportion electricity stealing behavior are developed, and the important reference value is provided for coping with the more severe electricity stealing prevention situation of the power grid company.
Disclosure of Invention
The invention provides a detection system and a detection method based on line loss and user power consumption covariance aiming at the scene of group fixed proportion electricity stealing, and aims to solve the problem of poor effect of the prior art method in the scene and better deal with the problem of electricity stealing prevention of a power grid company.
Briefly describing the technical scheme of the invention patent as follows:
with the increase of the number of electricity stealing users, after the meter electricity consumption vectors of the electricity stealing users subjected to standardization processing are superposed, the covariance between the line electricity loss vector of the transformer area subjected to standardization processing and the meter electricity consumption vector of the users is increased, and the covariance is called as incremental property; the invention converts the detection problem of group fixed proportion electricity stealing behavior into a combination optimization problem by utilizing the incremental property, and realizes the solution of the combination optimization problem, and the specific technical scheme is as follows:
a detection system adapted for group fixed proportion electricity stealing behavior, comprising: the device comprises a preprocessing module, a detection module and a judgment module;
the preprocessing module is connected with the detection module and carries out data transmission; the detection module is connected with the judgment module and carries out data transmission;
the preprocessing module is used for:
firstly, extracting electric quantity data of a user in unit time from an electric consumption information acquisition system, carrying out vectorization and standardization processing on data, and inputting the data into a detection module;
extracting the electric quantity data of the line loss of the distribution room in unit time from the integrated line loss management system, carrying out vectorization and standardization processing on the data, and inputting the data into a detection module;
the detection module is used for:
detecting the group fixed proportion electricity stealing behavior by using the increasing property of the covariance of the line loss quantity vector of the station area subjected to standardized processing and the meter power consumption quantity vector of the electricity stealing users with fixed proportion;
judging whether the power utilization scene is a fixed proportion power stealing scene;
output electricity stealing suspicion user set CsusAnd transmitting to the judging module;
the judging module is used for: and calculating the electricity stealing suspicion degree of the user.
A detection method for applying the detection system to group fixed proportion electricity stealing behavior comprises the following steps:
s1, the preprocessing module extracts the meter electricity consumption sequence of each user in the station area k in one day from the electricity information acquisition system to form the meter electricity consumption vector of the user i
Figure GDA0003515551400000031
Wherein the content of the first and second substances,
Figure GDA0003515551400000032
measuring the electricity consumption for the meter of the user i in the Tth time period;
s2, the preprocessing module extracts a line loss electric quantity sequence of the transformer area k in one day from the integrated line loss management system to form a line loss electric quantity vector omega of the transformer area kk＝[ωk,1,ωk,2,...,ωk,T]TWherein, ω isk,TLine loss capacity in the Tth time period of the transformer area k;
s3, using formula (1)
Figure GDA0003515551400000033
Performing standardization treatment to obtain
Figure GDA0003515551400000034
Figure GDA0003515551400000035
Using formula (2) to convert omegakPerforming standardization treatment to obtain
Figure GDA0003515551400000036
Figure GDA0003515551400000037
S4, mixing
Figure GDA0003515551400000038
And
Figure GDA0003515551400000039
as the input of the detection module, wherein A is the set formed by all users in the station zone k;
when the meter power consumption vectors of any two electricity stealing users with fixed proportion in the transformer area k are positively correlated, the covariance of the meter power consumption vectors of the standardized users and the line loss power vectors of the transformer area is increased along with the increase of the number of the electricity stealing users after the meter power consumption vectors of the standardized electricity stealing users are superposed, and the covariance is called as the incremental property;
the detection module detects the group fixed proportion electricity stealing behavior by utilizing the incremental property;
s5, the detection module considers that electricity stealing users in the distribution room k do not necessarily adopt a fixed proportion electricity stealing mode under the actual condition, and the detection module also comprises a fixed proportion electricity stealing scene judgment link. The detection module judges whether the power utilization scene is a fixed proportion power stealing scene or not and outputs a power stealing suspicion user set CsusNamely, a discrimination link of a fixed proportion electricity stealing scene;
s6, when the detection module operates each time, the detection module operates on the same day and in the same platform area
Figure GDA00035155514000000310
And
Figure GDA0003515551400000041
as input, with CsusAs an output; after the detection module processes the data of the station area k for M days, M output C are processedsusTransmitting the data to a judging module;
s7, the judging module counts the belongings of the user i to the user C in M dayssusNumber of times MiCalculating suspicion degree of electricity stealing γ of user i according to equation (3)i，
Figure GDA0003515551400000042
Gamma of user iiThe larger the electricity stealing suspicion, the gamma is selectediThe top several users are on-site audited. Because the electricity stealing behaviors of electricity stealing users in different transformer areas are different, the suspicion degrees of electricity stealing calculated in different transformer areas are different in distribution. Selecting a plurality of gamma according to actual specific conditions of each station areaiThe highest users are under inspection, for example: if the area AT belongs to a high power stealing area, then γ can be choseniTop 15 user audits; assuming that the cell BT belongs to a power stealing low power generation cell,then only γ needs to be selectediChecking the top 1-2 users.
On the basis of the technical scheme, the incremental property is as follows: when the meter power consumption vectors of any two electricity stealing users with fixed proportion in the transformer area k are positively correlated, the covariance of the line loss power vector of the transformer area k and the meter power consumption vector of the electricity stealing users with fixed proportion has the incremental property along with the increase of the number of the electricity stealing users with fixed proportion, as shown in formula (4),
Figure GDA0003515551400000043
wherein the content of the first and second substances,
Figure GDA0003515551400000044
represents: computing
Figure GDA0003515551400000045
And
Figure GDA0003515551400000046
the covariance of (a); c is the set of stolen users in zone k,
Figure GDA0003515551400000047
the normalized meter power vector for user j is counted,
Figure GDA0003515551400000048
calculating a power consumption vector for the normalized meter of the user h;
Figure GDA0003515551400000049
on the basis of the technical scheme, the detection module converts the detection problem of group fixed proportion electricity stealing behavior into a combination optimization problem shown in a formula (5),
Figure GDA0003515551400000051
Figure GDA0003515551400000052
wherein the content of the first and second substances,
Figure GDA0003515551400000053
represents: set P is any non-empty subset of set a.
On the basis of the technical scheme, the detection module decomposes the combinatorial optimization problem into a plurality of sub-problems, and each sub-problem is expressed as: the user combination which is matched with the user i and enables the covariance of the meter electricity consumption vector of the user and the line loss electricity vector of the station area k to be maximum is shown as a formula (6),
Figure GDA0003515551400000054
wherein the content of the first and second substances,
Figure GDA0003515551400000055
represents: piAny subset after user i is removed for set a.
On the basis of the technical scheme, a global optimal solution P is assumedmaxComprises the following steps: after the search of the detection module, the user combination with the maximum covariance of the line loss electric quantity vector of the distribution room and the meter electricity consumption vector of the user is formed, and the sum of the covariance of the line loss electric quantity vector of the distribution room and the meter electricity consumption vector of the user is PmaxSet of suspected users, P, as station area kmaxIs expressed by the formula (7),
Figure GDA0003515551400000056
the detection module searches the optimal solution P of each sub-problemimaxTo realize a global optimum solution PmaxAnd (4) solving.
On the basis of the technical scheme, the detection module realizes the global optimal solution PmaxThe solving method specifically comprises the following steps:
S4.1、first according to
Figure GDA0003515551400000057
Arrange users in descending order, search preferentially
Figure GDA0003515551400000058
P corresponding to larger userimax(ii) a Since the probability that the user with the larger covariance belongs to the electricity stealing user set is higher, the users are firstly ranked in a descending order according to the covariance, the optimal combination corresponding to the user with the higher probability is preferentially searched, and the global search can be accelerated to a certain extent.
S4.2, 1-order optimization: if it is used
Figure GDA00035155514000000610
Let i → j be the 1 st order incremental path of user i, and corresponding j is marked as j1(ii) a Remember that
Figure GDA0003515551400000061
J with the largest increment1Is j1maxThe corresponding incremental path i → j1maxCalled the 1 st order most increased path; the 1 st order optimization process is to find i → j1maxThe process of (2);
s4.3, 2-order optimization: at i → j1maxOn the basis of (i → j) is found1max→j2max；
S4.4, n-order optimization: at i → j1max→…→j(n-1)maxOn the basis, find i → j1max→…→j(n-1)max→jnmax；
S4.5, cutoff condition: after n-order optimization, if n is | A | -1, wherein | A | represents the number of elements in the set A, which indicates that the optimization process has finished traversing all users, the optimization stops; or after n-order optimization, when n + 1-order optimization is carried out, if an n + 1-order incremental path does not exist, the covariance of the line loss electric quantity vector of the transformer area k and the meter electricity consumption vector of the electricity stealing users with fixed proportion is not increased any more, and the optimization is terminated;
s4.6, after the optimization is ended, Pimax＝{j1max,j2max,…,jkmax}；
S4.7, for each user i, steps S4.1-S4.6 are taken to search for the corresponding PimaxGlobal optimal solution PmaxAs shown in the formula (8),
Figure GDA0003515551400000062
on the basis of the above technical solution, the specific steps of step S5 are:
correlation coefficient corresponding to global optimum solution
Figure GDA0003515551400000063
As a judgment index for whether the electricity stealing scene is a fixed proportion electricity stealing scene or not, wherein,
Figure GDA0003515551400000064
represents:
Figure GDA0003515551400000065
and
Figure GDA0003515551400000066
a pearson correlation metric of;
setting a Threshold when
Figure GDA0003515551400000067
Then, the analyzed power utilization scene is a fixed proportion power stealing scene (namely, the fixed proportion power stealing is taken as the main part), and the detection module outputs a suspected power stealing user set Csus＝Pmax(ii) a When in use
Figure GDA0003515551400000068
Then the analyzed power utilization scene is not a fixed proportion power stealing scene, and in order to avoid the failure of the algorithm, the output of the detection module
Figure GDA0003515551400000069
And a locking detection module.
The invention has the following beneficial technical effects:
(1) the invention only takes the power consumption of the user and the loss electric quantity of the transformer area (namely the line loss electric quantity) as the original data, does not need to add ammeter equipment for collecting the rest electric quantity on the premise of fully ensuring the privacy of the user, can be simultaneously applied to a special transformer area and a public transformer area, and has excellent practical conditions;
(2) aiming at the short board that the detection accuracy and the recall ratio of the current detection method are low in the group fixed proportion electricity stealing scene, the detection of the electricity stealing users is realized by establishing a mathematical model of the electricity stealing users in the group fixed proportion and the electricity loss (namely the line loss) of the distribution room in a way of solving a combined optimization problem, and meanwhile, the method has a deep theoretical basis and a superior engineering application effect;
(3) the method considers that the user does not necessarily adopt the electricity stealing mode with the fixed proportion, and is specially provided with the discrimination link of the electricity stealing scene with the fixed proportion, so that the failure problem caused by a non-target scene is avoided, and the model has higher robustness.
Drawings
The invention has the following drawings:
fig. 1 is a flow chart of the detection method for population fixed-proportion electricity stealing behavior according to the invention.
FIG. 2 is a schematic diagram of the detection module in the detection system of the present invention searching for the optimal solution of the subproblem.
Fig. 3 is a graph showing the recall ratio of 5 detection methods according to the second embodiment as a function of the number of electricity stealing users.
Fig. 4 is a curve diagram illustrating the accuracy of 5 detection methods according to the second embodiment as the number of power stealing users changes.
FIG. 5 is a schematic diagram of a framework of a system for detecting group electric larceny behavior with a fixed ratio according to the present invention
Detailed Description
For better understanding of the above technical solutions, the following detailed descriptions will be provided in conjunction with the drawings and the detailed description of the embodiments.
As shown in fig. 5, a detection system suitable for group fixed proportion electricity stealing behavior comprises: the device comprises a preprocessing module, a detection module and a judgment module;
the preprocessing module is connected with the detection module and carries out data transmission; the detection module is connected with the judgment module and carries out data transmission;
the preprocessing module is used for:
firstly, extracting electric quantity data of a user in unit time from an electricity utilization information acquisition system, carrying out vectorization and standardization processing on the data, namely standardizing an electric quantity vector, and inputting the data into a detection module;
extracting the electric quantity data of the line loss of the distribution room in unit time from the integrated line loss management system, carrying out vectorization and standardization processing on the data, and inputting the data into a detection module; that is, as shown in fig. 5, the raw power consumption data is vectorized, and the line loss vector is calculated and normalized.
The detection module is used for:
the method comprises the following steps of firstly, detecting group fixed proportion electricity stealing behaviors by utilizing the incremental property of the covariance of line loss vectors of a platform area subjected to standardized processing and meter metering electricity consumption vectors of electricity stealing users with fixed proportion along with the increase of the number of the electricity stealing users with fixed proportion;
judging whether the power utilization scene is a fixed proportion power stealing scene;
output electricity stealing suspicion user set CsusAnd transmitting to the judging module;
namely, LLC (detection method suitable for group fixed proportion electricity stealing behavior) algorithm detection is carried out, and electricity stealing suspicion user set C of each day d is outputsus。
The judging module is used for: calculating suspicion of electricity stealing of users, i.e. calculating gamma of each user iiSelecting gammaiThe top several users are on-site audited.
The technical solution of the present invention is further described in detail with reference to fig. 1 and 2. FIG. 1 is a schematic flow chart of the detection method for population fixed-proportion electricity stealing behavior of the invention, which mainly comprises the following steps:
s1, data extraction and vectorization
The preprocessing module extracts the meter power consumption sequence of each user in the station area k in one day from the power consumption information acquisition system to form a meter power consumption vector of the user i
Figure GDA0003515551400000091
S2, the preprocessing module extracts a line loss electric quantity sequence of the transformer area k in one day from the integrated line loss management system to form a line loss electric quantity vector omega of the transformer area kk＝[ωk,1,ωk,2,...,ωk,T]T；
S3, data standardization processing
By using the formula (1)
Figure GDA0003515551400000092
Performing standardization to obtain
Figure GDA0003515551400000093
Figure GDA0003515551400000094
Using formula (2) to convert omegakPerforming standardization treatment to obtain
Figure GDA0003515551400000095
Figure GDA0003515551400000096
S4, detection module will obtain
Figure GDA0003515551400000097
And
Figure GDA0003515551400000098
for input, outputting the electricity stealing suspicion user set of the station area k in a mode of solving a combined optimization problemCsusSpecifically, as follows,
1) search of sub-problem optimal solution
When the meter power consumption vectors of any two electricity stealing users with fixed proportion in the station area k are positively correlated, searching for a user combination P which is matched with the user i and enables the maximum covariance of the meter power consumption vectors of the users and the line loss power vectors of the station area k based on the incremental property (the incremental property of covariance) of the covariance of the meter power consumption vectors of the electricity stealing users with fixed proportion and the line loss power vectors of the station area k along with the increase of the number of the electricity stealing users with fixed proportionimax；
2) Search of global optimal solution
Searching P corresponding to each user iimaxWherein P is the largest covariance of the electricity consumption vector of the meter of the user and the line loss electricity vector of the station area kimaxIs the global optimal solution Pmax；
S5, judgment of target scene
Calculating a correlation coefficient corresponding to the global optimal solution
Figure GDA0003515551400000099
Setting a Threshold; when in use
Figure GDA00035155514000000910
Meanwhile, the analyzed scene is considered to be mainly based on electricity stealing in a fixed proportion, and the detection module outputs an electricity stealing suspected user set Csus＝Pmax(ii) a When in use
Figure GDA00035155514000000911
And (4) considering that the analysis object is not the target scene, and outputting to avoid algorithm failure
Figure GDA0003515551400000101
A lockout detection module;
s6, after the detection module processes the data of the station area k for M days, the C of each day is processedsusTransmitting the data to a judging module;
s7, judging whether the user i belongs to C within M dayssusNumber of times MiTo calculate the suspicion degree gamma of electricity stealing of the user ii(ii) a And finally, selecting a plurality of users with the highest electricity stealing suspicion degree to carry out field inspection.
Fig. 2 is a schematic diagram of the principle of searching the optimal solution of the subproblem in the detection module 201, and the specific steps include:
s4.1, first according to
Figure GDA0003515551400000102
Sort users down, search first
Figure GDA0003515551400000103
P corresponding to larger userimax。
S4.2, 1-order optimization: if it is not
Figure GDA0003515551400000104
Let i → j be the 1 st order incremental path of user i, and corresponding j is marked as j1(ii) a Remember that
Figure GDA0003515551400000105
J with the largest increment1Is j1maxThe corresponding incremental path i → j1maxCalled the 1 st order most increased path; the 1 st order optimization process is to find i → j1maxThe process of (1).
S4.3, 2-order optimization: at i → j1maxOn the basis of (i → j) is found1max→j2max。
S4.4) n-order optimization: at i → j1max→…→j(n-1)maxOn the basis, find i → j1max→…→j(n-1)max→jnmax。
S4.5, cutoff condition: after n-order optimization, if n is | A | -1, the optimization process is described to have traversed all users, and the optimization is stopped; or after the n-order optimization, when the n + 1-order optimization is carried out, if the n + 1-order incremental path does not exist, the covariance of the line loss electric quantity vector of the station area k and the meter electricity consumption vector of the electricity stealing users with fixed proportion is not increased any more, and the optimization is terminated.
S4.6, optimizing and terminatingThen, Pimax＝{j1max,j2max,…,jkmax}。
Examples
The following description will explain the effect of the present invention by referring to two examples. The power utilization data provided by a certain provincial company of the national power grid is used as an original data set, and the information of the data set is shown in table 1. All users in the data set come from the station area with the line loss rate lower than 3%, and the data set can be considered as normal users. The method comprises the steps of selecting 500 three-phase users as normal power utilization data in 61 days from 8 months in 2019 to 9 months in 2019, and carrying out power stealing transformation on the normal power utilization data by depending on a power stealing simulation experiment platform of an anti-power stealing center of the Chinese power science research institute. The detection accuracy is evaluated by adopting two indexes of recall ratio and precision ratio, the application effect of the invention is illustrated by comparing the other 4 electricity stealing detection methods, and the introduction of the 4 comparison methods is shown in table 2.
TABLE 1 original data set description Table
Figure GDA0003515551400000111
TABLE 2 description of the comparative testing method
Figure GDA0003515551400000112
The first embodiment is as follows:
selecting 61 days of data, dividing 500 three-phase users into 10 groups representing 10 power supply districts, selecting 6 users in each district, and carrying out fixed proportion electricity stealing transformation, wherein the proportion of electricity stealing users in each district is 12%. 100 repeated experiments were performed by random configuration of the area and random selection by the electricity stealing users.
Table 3 shows the average values of the recall ratio and the precision ratio of the 5 methods in 100 experiments, and it can be seen that the recall ratio and the precision ratio of the invention are both above 0.95 and much higher than the indexes of CFSFDP and LOF, and the indexes are improved by nearly 25% on the basis of PPC (0.787,0.758) and MIC (0.768,0.737), which indicates that the invention has superior detection accuracy in a group fixed proportion electricity stealing scene.
Table 4 shows the standard deviations of the recall ratio and the precision ratio of the 5 methods in 100 experiments, and it can be seen that the standard deviations of the recall ratio of the 5 methods are not very different and are all between 0.04 and 0.06; the standard deviation of the precision of the 5 methods is distributed between 0.09 and 0.19, wherein the standard deviation of the precision of the invention is 0.103 and is only higher than the standard deviation of LOF, which shows that the invention has good operation stability in practical application.
TABLE 3 table of average values of recall ratio and precision ratio of 5 detection methods in the first embodiment
Figure GDA0003515551400000121
TABLE 4 Standard deviations of recall and precision for 5 detection methods of example one
Figure GDA0003515551400000122
Example two:
selecting data of 61 days, dividing 500 three-phase users into 10 groups, representing 10 power supply districts, increasing the number of electricity stealing users from 2 to 16 (step length is 2) under the condition of ensuring that the total number of the users in each district is not changed, and explaining the influence of the change of the number of the electricity stealing users with fixed proportion on different detection methods.
Fig. 3 shows the change of recall ratio of 5 methods in the process, and fig. 4 shows the change of precision ratio. As can be seen from fig. 3 and 4, when the number of electricity stealing users is small, the PCC, the MIC and the present invention have high detection accuracy, but as the number of electricity stealing users increases, the recall ratio of the PCC and the MIC decreases rapidly and is gradually separated from the index of the present invention; in the whole process, the reduction of the recall ratio of PCC and MIC is respectively 0.231(0.92-0.689) and 0.251(0.911-0.654), while the invention only corresponds to 0.111(0.998-0.887), and the precision ratios of the PCC and the MIC also have similar changes: this indicates that: when a group fixed proportion electricity stealing behavior occurs, negative effects are brought to the detection accuracy of PCC and MIC, and the negative effects are further aggravated as the scale of electricity stealing users increases; the indexes of the invention are reduced, but the indexes are kept at a high score, which shows that the invention has good operation robustness in the scene.
In summary, the group fixed proportion electricity stealing behavior detection method provided by the invention fully utilizes the mathematical characteristics between the electricity quantity of the group fixed proportion electricity stealing users and the line loss of the distribution room, can realize accurate check and complete check of the group fixed proportion electricity stealing users on the premise of fully respecting the privacy of the users, and simultaneously adds a target scene judgment link, so that the model has higher operation stability and robustness.
It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention. Thus, it is intended that the present invention cover the modifications and variations of this invention provided they come within the scope of the appended claims and their equivalents.
Those not described in detail in this specification are within the knowledge of those skilled in the art.

Claims (8)

1. A system for detecting group-based fixed-proportion electricity stealing behavior, comprising: the device comprises a preprocessing module, a detection module and a judgment module;
the preprocessing module is connected with the detection module and carries out data transmission; the detection module is connected with the judgment module and carries out data transmission;
the preprocessing module is used for:
firstly, extracting electric quantity data of a user in unit time from an electric consumption information acquisition system, carrying out vectorization and standardization processing on data, and inputting the data into a detection module;
extracting the electric quantity data of the distribution room line loss in unit time from the integrated line loss management system, carrying out vectorization and standardization processing on the data, and inputting the data into a detection module;
the detection module is used for:
detecting the group fixed proportion electricity stealing behavior by using the increasing property of the covariance of the line loss quantity vector of the station area subjected to standardized processing and the meter power consumption quantity vector of the electricity stealing users with fixed proportion;
judging whether the power utilization scene is a fixed proportion power stealing scene;
output electricity stealing suspicion user set CsusAnd transmitting to the judging module;
the judging module is used for: and calculating the electricity stealing suspicion degree of the user.
2. A method for detecting group fixed-proportion electricity stealing behavior by using the detection system of claim 1, which comprises the following steps:
s1, the preprocessing module extracts the meter electricity consumption sequence of each user in the station area k in one day from the electricity information acquisition system to form the meter electricity consumption vector of the user i
Figure FDA0003515551390000011
Wherein the content of the first and second substances,
Figure FDA0003515551390000012
measuring the electricity consumption for the meter of the user i in the Tth time period;
s2, the preprocessing module extracts a line loss electric quantity sequence of the transformer area k in one day from the integrated line loss management system to form a line loss electric quantity vector omega of the transformer area kk＝[ωk,1,ωk,2,...,ωk,T]TWherein, ω isk,TLine loss capacity in the Tth time period of the transformer area k;
s3, using formula (1)
Figure FDA0003515551390000021
Performing standardization treatment to obtain
Figure FDA0003515551390000022
Figure FDA0003515551390000023
Using formula (2) to convert omega to omegakPerforming standardization treatment to obtain
Figure FDA0003515551390000024
Figure FDA0003515551390000025
S4, mixing
Figure FDA0003515551390000026
And
Figure FDA0003515551390000027
as the input of the detection module, wherein A is the set formed by all users in the station zone k;
when the meter power consumption vectors of any two electricity stealing users with fixed proportion in the transformer area k are positively correlated, the covariance of the meter power consumption vectors of the standardized users and the line loss power vectors of the transformer area is increased along with the increase of the number of the electricity stealing users after the meter power consumption vectors of the standardized electricity stealing users are superposed, and the covariance is called as the incremental property;
the detection module detects the group fixed proportion electricity stealing behavior by utilizing the incremental property;
s5, the detection module judges whether the power utilization scene is a fixed proportion power stealing scene or not, and outputs a power stealing suspicion user set Csus；
S6, when the detection module runs each time, the detection module operates in the same area on the same day
Figure FDA0003515551390000028
And
Figure FDA0003515551390000029
as input, with CsusAs an output; after the detection module processes the data of the station area k for M days, M output C are processedsusTransmitting the data to a judging module;
s7, the judging module counts the belongings of the user i to the user C in M dayssusNumber of times MiCalculating suspicion degree of electricity stealing γ of user i according to equation (3)i，
Figure FDA00035155513900000210
Gamma of user iiThe larger the electricity stealing suspicion is, the more gamma is selectediThe top several users are on-site audited.
3. A method of detecting group-wise fixed proportion electricity stealing behavior as claimed in claim 2, wherein: the incremental properties are: when the meter power consumption vectors of any two electricity stealing users with fixed proportion in the transformer area k are positively correlated, the covariance of the line loss power vector of the transformer area k and the meter power consumption vector of the electricity stealing users with fixed proportion has the incremental property along with the increase of the number of the electricity stealing users with fixed proportion, as shown in formula (4),
Figure FDA0003515551390000031
wherein the content of the first and second substances,
Figure FDA0003515551390000032
represents: calculating out
Figure FDA0003515551390000033
And
Figure FDA0003515551390000034
the covariance of (a); c is the set of stolen users in zone k,
Figure FDA0003515551390000035
the normalized meter power vector for user j is counted,
Figure FDA0003515551390000036
calculating a power consumption vector for the normalized meter of the user h;
Figure FDA0003515551390000037
4. a method of detecting group-specific proportion electricity stealing behavior as claimed in claim 3, wherein: the detection module converts the detection problem of group fixed proportion electricity stealing behavior into a combination optimization problem shown in a formula (5),
Figure FDA0003515551390000038
Figure FDA0003515551390000039
wherein, the first and the second end of the pipe are connected with each other,
Figure FDA00035155513900000310
represents: set P is any non-empty subset of set a.
5. The method for detecting group fixed-proportion electricity stealing behavior according to claim 4, wherein: the detection module decomposes the combinatorial optimization problem into a plurality of sub-problems, each sub-problem is expressed as: the user combination which is matched with the user i and enables the covariance of the meter electricity consumption vector of the user and the line loss electricity vector of the station area k to be maximum is shown as a formula (6),
Figure FDA00035155513900000311
wherein the content of the first and second substances,
Figure FDA00035155513900000312
represents: piAny subset after user i is removed for set a.
6. The method for detecting group fixed-proportion electricity stealing behavior according to claim 5, wherein: assuming a global optimal solution PmaxComprises the following steps: after the search of the detection module, the user combination with the maximum covariance of the line loss electric quantity vector of the station area and the electric quantity vector used for the user meter is calculated by PmaxSet of suspected users, P, as station area kmaxIs expressed by the formula (7),
Figure FDA0003515551390000041
the detection module searches the optimal solution P of each sub-problemimaxTo achieve a global optimal solution PmaxAnd (4) solving.
7. The method for detecting group fixed-proportion electricity stealing behavior according to claim 6, wherein: the detection module realizes the global optimal solution PmaxThe solving method specifically comprises the following steps:
s4.1, first according to
Figure FDA0003515551390000042
Arrange users in descending order, search preferentially
Figure FDA0003515551390000043
P corresponding to larger userimax；
S4.2, 1-order optimization: if it is not
Figure FDA0003515551390000044
Let i → j be the 1 st order incremental path of user i, and corresponding j is marked as j1(ii) a Remember that
Figure FDA0003515551390000045
J with the largest increment1Is j1maxThe corresponding incremental path i → j1maxCalled the 1 st order most increased path; the 1 st order optimization process is to find i → j1maxThe process of (2);
s4.3, 2-order optimization: at i → j1maxOn the basis of (i → j) is found1max→j2max；
S4.4, n-order optimization: at i → j1max→…→j(n-1)maxOn the basis, find i → j1max→…→j(n-1)max→jnmax；
S4.5, cutoff condition: after n-order optimization, if n ═ a | -1, the optimization stops, where | a | represents the number of elements in set a; or after n-order optimization, when n + 1-order optimization is carried out, if no n + 1-order incremental path exists, the optimization is terminated;
s4.6, after the optimization is ended, Pimax＝{j1max,j2max,…,jkmax}；
S4.7, for each user i, steps S4.1-S4.6 are taken to search for the corresponding PimaxGlobal optimal solution PmaxAs shown in the formula (8),
Figure FDA0003515551390000046
8. the method for detecting group fixed-proportion electricity stealing behavior of claim 7, wherein: the specific steps of step S5 are:
correlation coefficient corresponding to global optimum solution
Figure FDA0003515551390000051
As a judgment index for whether the electricity stealing scene is a fixed proportion electricity stealing scene or not, wherein,
Figure FDA0003515551390000052
represents:
Figure FDA0003515551390000053
and
Figure FDA0003515551390000054
a pearson correlation metric of;
setting a Threshold when
Figure FDA0003515551390000055
Then, the analyzed power utilization scene is a fixed proportion power stealing scene (namely, the fixed proportion power stealing is taken as the main part), and the detection module outputs a suspected power stealing user set Csus＝Pmax(ii) a When in use
Figure FDA0003515551390000056
The analyzed power utilization scene is not a fixed proportion electricity stealing scene, and the output of the detection module
Figure FDA0003515551390000057
And a locking detection module.
CN202110237945.5A 2021-03-04 2021-03-04 Detection system and detection method suitable for group fixed proportion electricity stealing behavior Active CN113514695B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110237945.5A CN113514695B (en) 2021-03-04 2021-03-04 Detection system and detection method suitable for group fixed proportion electricity stealing behavior

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110237945.5A CN113514695B (en) 2021-03-04 2021-03-04 Detection system and detection method suitable for group fixed proportion electricity stealing behavior

Publications (2)

Publication Number Publication Date
CN113514695A CN113514695A (en) 2021-10-19
CN113514695B true CN113514695B (en) 2022-05-06

Family

ID=78060921

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110237945.5A Active CN113514695B (en) 2021-03-04 2021-03-04 Detection system and detection method suitable for group fixed proportion electricity stealing behavior

Country Status (1)

Country Link
CN (1) CN113514695B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103208091A (en) * 2013-04-25 2013-07-17 国家电网公司 Electric larceny preventing method based on data mining of electric load management system
CN106066423A (en) * 2016-05-25 2016-11-02 上海博英信息科技有限公司 A kind of analysis method of opposing electricity-stealing based on Loss allocation suspicion analysis
KR101863326B1 (en) * 2018-02-07 2018-07-05 이엔티코리아 주식회사 System for preventing electricity theft
CN110264015A (en) * 2019-06-28 2019-09-20 国网河南省电力公司电力科学研究院 It opposes electricity-stealing and checks monitoring method and platform
CN110824270A (en) * 2019-10-09 2020-02-21 中国电力科学研究院有限公司 Electricity stealing user identification method and device combining transformer area line loss and abnormal events
CN111507611A (en) * 2020-04-15 2020-08-07 北京中电普华信息技术有限公司 Method and system for determining electricity stealing suspected user
CN111521868A (en) * 2020-04-28 2020-08-11 广东电网有限责任公司梅州供电局 Method and device for screening electricity stealing users based on big metering data

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9305448B2 (en) * 2014-04-04 2016-04-05 Sahibzada Ali Mahmud Securing distribution lines from pilferages
US20160161539A1 (en) * 2014-12-09 2016-06-09 Powerhive, Inc. Electricity theft detection system

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103208091A (en) * 2013-04-25 2013-07-17 国家电网公司 Electric larceny preventing method based on data mining of electric load management system
CN106066423A (en) * 2016-05-25 2016-11-02 上海博英信息科技有限公司 A kind of analysis method of opposing electricity-stealing based on Loss allocation suspicion analysis
KR101863326B1 (en) * 2018-02-07 2018-07-05 이엔티코리아 주식회사 System for preventing electricity theft
CN110264015A (en) * 2019-06-28 2019-09-20 国网河南省电力公司电力科学研究院 It opposes electricity-stealing and checks monitoring method and platform
CN110824270A (en) * 2019-10-09 2020-02-21 中国电力科学研究院有限公司 Electricity stealing user identification method and device combining transformer area line loss and abnormal events
CN111507611A (en) * 2020-04-15 2020-08-07 北京中电普华信息技术有限公司 Method and system for determining electricity stealing suspected user
CN111521868A (en) * 2020-04-28 2020-08-11 广东电网有限责任公司梅州供电局 Method and device for screening electricity stealing users based on big metering data

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Kedi Zheng ; Yi Wang ; Qixin Chen ; Yuanpeng Li.Electricity theft detecting based on density-clustering method.《2017 IEEE Innovative Smart Grid Technologies - Asia (ISGT-Asia)》.2017,全文. *
Yining Yang ; Runan Song ; Yang Xue ; Penghe Zhang ; Yuejie Xu ; Jinping K.A Detection Method for Group Fixed Ratio Electricity Thieves Based on Correlation Analysis of Non-Technical Loss.《IEEE Access》.2022, *
基于用电信息采集大数据的防窃电方法研究;窦健，刘宣，卢继哲，吴迪，王学伟;《电测与仪表》;20180110;全文 *

Also Published As

Publication number Publication date
CN113514695A (en) 2021-10-19

Similar Documents

Publication Publication Date Title
CN103810637B (en) motor vehicle insurance fraud detection method and system
CN109784388A (en) Stealing user identification method and device
CN110609200B (en) Power distribution network earth fault protection method based on fuzzy metric fusion criterion
CN108629525A (en) It is a kind of to consider that Severity method temporarily drops in the node voltage of load significance level
CN112308124B (en) Intelligent electricity larceny prevention method for electricity consumption information acquisition system
Yang et al. An intelligent singular value diagnostic method for concrete dam deformation monitoring
CN109521304A (en) Description and appraisal procedure and the device of resistance characteristics temporarily drop in a kind of node voltage
CN110244099A (en) Stealing detection method based on user's voltage
CN110390440B (en) Clustering and deep neural network-based intelligent ammeter user aggregate load prediction method
Qu et al. A residual convolutional neural network with multi-block for appliance recognition in non-intrusive load identification
CN113514695B (en) Detection system and detection method suitable for group fixed proportion electricity stealing behavior
CN112365164B (en) Energy characteristic portrait method for medium and large energy users based on improved density peak value rapid search clustering algorithm
Sun et al. Electricity theft detection method based on ensemble learning and prototype learning
CN111008673A (en) Method for collecting and extracting malignant data chain in power distribution network information physical system
Zhang et al. User power interaction behavior clustering analysis that is based on the self-organizing-center K-means algorithm
CN106442830A (en) Method and system for detecting alarm value of gas content of transformer oil
Sing et al. Local outlier factor based data mining model for three phase transmission lines faults identification
CN109887552A (en) A kind of water bursting source differentiation prediction technique
CN114066219A (en) Electricity stealing analysis method for intelligently identifying electricity utilization abnormal points under incidence matrix
CN112804197B (en) Power network malicious attack detection method and system based on data recovery
CN114897064A (en) Abnormal electricity consumption behavior detection device based on BiGAN and SVDD
CN114970693A (en) Charging pile user portrait method based on federal learning
CN106816871A (en) A kind of POWER SYSTEM STATE similarity analysis method
Jie et al. The study for data mining of distribution network based on particle swarm optimization with clustering algorithm method
CN117540277B (en) Lost circulation early warning method based on WGAN-GP-TabNet algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant