CN116910596A - User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering - Google Patents
User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering Download PDFInfo
- Publication number
- CN116910596A CN116910596A CN202310924145.XA CN202310924145A CN116910596A CN 116910596 A CN116910596 A CN 116910596A CN 202310924145 A CN202310924145 A CN 202310924145A CN 116910596 A CN116910596 A CN 116910596A
- Authority
- CN
- China
- Prior art keywords
- cluster
- user
- abnormal
- electricity
- month
- 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.)
- Granted
Links
- 230000005611 electricity Effects 0.000 title claims abstract description 127
- 238000004458 analytical method Methods 0.000 title claims abstract description 31
- 230000002159 abnormal effect Effects 0.000 claims abstract description 80
- 238000012549 training Methods 0.000 claims abstract description 30
- 230000002265 prevention Effects 0.000 claims abstract description 8
- 238000004138 cluster model Methods 0.000 claims abstract 2
- 238000004422 calculation algorithm Methods 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 3
- 235000014435 Mentha Nutrition 0.000 claims description 2
- 241001072983 Mentha Species 0.000 claims description 2
- 235000014569 mints Nutrition 0.000 claims description 2
- 230000008901 benefit Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 208000013935 Electric injury Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a user electricity stealing analysis method, equipment and a storage medium based on improved DBSCAN clustering, which are used for acquiring lost electric quantity of an abnormal platform region, power consumption related data of users under the abnormal platform region and related data of typical electricity stealing users in an electricity stealing prevention sample library; calculating a training sample data set of a user under an abnormal platform area, wherein the training sample data set of a typical electricity stealing user; according to the training sample data set of the user under the abnormal platform region, the training sample data set of the typical electricity stealing user obtains a cluster of the user under the abnormal platform region and the typical electricity stealing user through an improved DBSCAN cluster model; acquiring effective clusters in the cluster according to the average value of the contour coefficients; and acquiring intersection of the users in the effective clusters and the users under the abnormal area to obtain suspected electricity stealing users under the abnormal area. The invention effectively locates suspected users of electricity theft, improves working efficiency and ensures benefits of power supply companies.
Description
Technical Field
The invention relates to a user electricity stealing analysis method, equipment and a storage medium based on improved DBSCAN clustering, and belongs to the technical field of electricity utilization detection analysis of an electric power system.
Background
With the rapid promotion of market economy and the daily and monthly variation of science and technology, the electricity stealing phenomenon is increasingly vigorous and the means is higher while the demand of society for electric energy is increased. The electricity utilization clients steal electricity for the purposes of reducing the metering electric energy and paying electric fees, so that not only is economic loss caused to power supply companies, but also the normal power supply and utilization order is disturbed.
The electric power is an energy source with high risk, the electricity stealing behavior has great potential safety hazard, fire, clicking and electric injury are easy to be caused, the personal and property safety of people is threatened, and unstable factors are brought to the harmony of society. For the power grid, the power stealing behavior increases the line loss, so that the abnormal large change amplitude of the line loss rate can cause various abnormal phenomena in the power supply station area, thereby not only damaging the power infrastructure of the power grid, but also threatening the safe and stable operation of the power grid, and simultaneously preventing the development of the intelligent power grid.
The traditional electricity larceny detection mainly depends on modules such as metering operation monitoring and line loss analysis, and manual analysis is needed by technicians with abundant experience, but the methods are low in accuracy, large in workload and low in efficiency. In the electricity big data age, mass data accumulated by a new generation electricity information acquisition system and an electricity marketing system can be utilized, and electricity stealing analysis can be performed through a big data algorithm or a clustering algorithm. However, the conventional DBSCAN clustering algorithm cannot meet the analysis requirement of mass data, because it performs a large amount of global computation by randomly selecting core points, resulting in large computation amount and low analysis efficiency.
Therefore, the technical problem of the traditional DBSCAN algorithm for carrying out electricity stealing analysis on a large amount of electricity consumption data is urgent to be solved by the person skilled in the art.
Disclosure of Invention
The purpose is as follows: in order to overcome the defects in the prior art, the invention provides the user electricity larceny analysis method based on improved DBSCAN (clustering algorithm based on density) clustering, which utilizes mass electricity utilization client data accumulated by a new generation electricity utilization information acquisition system and an electric power marketing system, effectively locates suspected electricity larceny users through a big data analysis means, improves the working efficiency and ensures the benefits of power supply companies.
The technical scheme is as follows: in order to solve the technical problems, the invention adopts the following technical scheme:
in a first aspect, a method for user electricity theft analysis based on improved DBSCAN clustering includes the steps of:
step 1: and acquiring the lost electric quantity of the abnormal area, the power consumption related data of the users under the abnormal area and the related data of the typical power stealing users in the power stealing prevention sample library.
Step 2: and calculating a training sample data set of the user under the abnormal area according to the lost electric quantity of the abnormal area, the related data of the power consumption of the user under the abnormal area and the related data of the typical power stealing user in the anti-power stealing sample library.
Step 3: and obtaining clustering clusters of the users under the abnormal area and the typical electricity stealing users through an improved DBSCAN clustering model according to the training sample data set of the users under the abnormal area and the training sample data set of the typical electricity stealing users.
Step 4: and calculating the average value of the contour coefficients of each cluster in the cluster, and acquiring the effective clusters in the cluster according to the average value of the contour coefficients.
Step 5: and acquiring intersection of the users in the effective clusters and the users under the abnormal area to obtain suspected electricity stealing users under the abnormal area.
Preferably, the step 1 specifically includes:
step 1.1: and acquiring the line loss rate of each station area, defining the station area as an abnormal station area when the line loss rate of the station area exceeds the theoretical line loss rate, and acquiring the month loss electric quantity of the abnormal station area.
Step 1.2: and acquiring the month electricity consumption, the month trend of the electric quantity, the month average voltage, the month average current and the month average active power of the users under the abnormal platform area.
Step 1.3: and acquiring the month electricity consumption, the month electricity tendency, the month average voltage, the month average current and the month average active power of a typical electricity stealing user in an electricity stealing prevention sample library.
Preferably, the step 2 specifically includes:
step 2.1: obtaining correlation coefficient of month electricity consumption of users under abnormal areas and month electricity loss of abnormal areasWherein (1)>And the power consumption related coefficient of the user in the ith abnormal area is represented, i epsilon 1,2,3 … …, m and m are the number of the users in the abnormal area.
Step 2.1: obtaining the correlation coefficient of the month electricity consumption of a typical electricity stealing user and the month lost electricity of an abnormal platform areaWherein (1)>The j-th typical electricity stealing user power consumption correlation coefficient is represented, j epsilon 1,2,3 … …, k and k are the number of typical electricity stealing users.
Step 2.3: taking the electricity month trend of the user in the ith abnormal area as a characteristicCalculating the month average voltage, month average current and month average active power of the user under the ith abnormal platform area to obtain the characteristic ∈10->Will->Feature vector for composing user under ith abnormal area +.>Acquiring feature vectors of all users under the abnormal platform region to form a training sample data set D of the users under the abnormal platform region m ，/>
Step 2.4: taking the month trend of the electricity quantity of the j-th typical electricity stealing user as a characteristicCalculating month average voltage, month average current and month average active power of the j-th typical electricity stealing user to obtain characteristic +.>Will->Characteristic vector +.>Acquiring feature vectors of all typical electricity stealing users to form a training sample data set D of the typical electricity stealing users k ，/>
Preferably, the step 3 specifically includes:
step 3.1: training sample data set D of user under abnormal platform area m As a DBSCAN clustering sample set, clustering the DBSCAN clustering sample set by DBSCAN to obtain a clustering result, and taking boundary points and noise points in the clustering result as an abnormal user set D of preliminary screening l 。
Step 3.2: selecting training sample data set D of typical electricity stealing users k As a core object, a core object Ω=d is initialized k 。
Step 3.3: the method comprises the following steps of finding out density reachable objects with the number of samples being greater than or equal to a number threshold MinPts in a core object omega neighborhood radius epsilon through iteration. Initializing the cluster number k=0 of the clusters, initializing the unvisited sample set Γ=d, d=d k ∪D l Initializing cluster clusters Is null.
Step 3.4: judging whether the core object omega is null value, if soThe algorithm ends and cluster c= { C is output 1 ,C 2 ,C 3 ……C j -a }; otherwise, step 3.5 is entered.
Step 3.5: selecting a first core object P from a core object set omega, calculating density reachable points in a neighborhood radius epsilon of the core object P, initializing a set N= { P } of a current core object cluster, initializing a class sequence number j=j+1, updating a non-accessed sample set Γ=Γ - { P }, and initializing a current cluster sample set C j ＝{P}。
Step 3.6: judging whether the set N of the current core object cluster is empty, if soThen the current cluster C k After the completion of the generation, the cluster division set c= { C is updated 1 ,C 2 ,C 3 ……C j Update core object set Ω=Ω -C j And step 3.4; otherwise, step 3.7 is entered.
Step 3.7: randomly selecting a core object P' from a set N of current core object clusters by the neighbor domainThe sample number threshold value Minpts finds out all the sub-sample sets N '= { P' } in the epsilon neighborhood, and updates the current cluster sample set C j ＝C j And updating the unvisited sample set Γ=Γ -N ', updating the current cluster core object set N=N- (N '. Cndot. OMEGA) -P ', and entering step 3.6.
Preferably, the step 4 specifically includes:
step 4.1: and calculating the average distance a between each sample point P in each cluster in the cluster C and other sample points in the same cluster.
Step 4.2: and calculating the average distance b between each sample point P in each cluster in the cluster C and all sample points in other clusters.
Step 4.3: the contour coefficient S of each sample point P in each cluster in the cluster C is calculated, and the calculation formula of the contour coefficient is as follows:
step 4.4: calculating average value of contour coefficients of all sample points in each cluster in cluster C
Step 4.5: when (when)When the threshold delta is larger than the set threshold delta, the corresponding cluster in the cluster clusters is an effective cluster.
Preferably, the delta is set to 0.8.
In a second aspect, a user electricity theft analysis device based on improved DBSCAN clustering includes a processor and a storage medium.
The storage medium is for storing instructions.
The processor is operative in accordance with instructions to perform the steps of the user electricity theft analysis method of the first aspect based on improved DBSCAN clustering.
In a third aspect, a computer readable storage medium has stored thereon a computer program which when executed by a processor implements the steps of the user electricity theft analysis method of the first aspect based on improved DBSCAN clustering.
The beneficial effects are that: according to the user electricity stealing analysis method, the device and the storage medium based on the improved DBSCAN clustering, important data related to electricity stealing are selected, after data preprocessing, clustering advantages of the DBSCAN clustering algorithm are fully utilized, abnormal users are primarily screened out, the primarily screened abnormal users and typical electricity stealing user sample data are used as sample sets, the sample set data are clustered through the improved DBSCAN clustering algorithm, and therefore abnormal data are clustered, suspected electricity stealing users are accurately identified according to clustering results and profile coefficients, suspected electricity stealing users are effectively located, working efficiency is improved, and benefits of power supply companies are guaranteed.
Drawings
FIG. 1 is a flow chart of a user electricity larceny analysis method of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made more apparent and fully by reference to the accompanying drawings, in which embodiments of the invention are shown, and in which it is evident that the embodiments shown are only some, but not all embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without any inventive effort, are intended to be within the scope of the present invention.
The invention will be further described with reference to specific examples.
Example 1:
as shown in fig. 1, this embodiment introduces a user electricity stealing analysis method based on improved DBSCAN clustering, which specifically includes the following steps:
step 1: and acquiring the lost electric quantity of the abnormal area, the power consumption related data of the users under the abnormal area and the related data of the typical power stealing users in the power stealing prevention sample library. The method specifically comprises the following steps:
step 1.1: and acquiring the line loss rate of each area based on the line loss module of the electricity consumption information acquisition system, defining the area as an abnormal area when the line loss rate of the area exceeds the theoretical line loss rate, and acquiring the month loss electricity quantity of the abnormal area.
Step 1.2: and acquiring data of the monthly power consumption, the monthly trend of the electric quantity, the monthly average voltage, the monthly average current and the monthly average active power of the users under the abnormal platform area.
Step 1.3: and acquiring month electricity consumption, month trend of electricity quantity, month average voltage, month average current and month average active power data of typical electricity stealing users in an electricity stealing prevention sample library.
Step 2: and calculating the Pearson correlation coefficient of the electricity consumption of the user under the abnormal area and the electricity loss of the area, selecting the electricity stealing characteristic value according to the electricity consumption related data of the user under the abnormal area and the electricity consumption related data of the typical electricity stealing user, and constructing a sample training set according to the Pearson correlation coefficient and the electricity stealing characteristic value. The method specifically comprises the following steps:
step 2.1: according to a Pearson correlation coefficient calculation formula, calculating the Pearson correlation coefficient rho of the user power consumption and the station loss power under the abnormal station and the typical power stealing user power consumption and the station loss power respectively X,Y ：
Wherein ρ is the pearson correlation coefficient, X is the user power consumption, Y is the station area loss power, cov (X, Y) is the covariance of the variable X and the variable Y, σ X And sigma (sigma) Y Is the mean square error.
The covariance and standard deviation between the variables are calculated to obtain pearson correlation coefficients, namely:
wherein X is i For the user's i day power consumption, Y i The power is lost for the ith land area,average value of electricity consumption of users respectivelyAverage value of electric quantity lost in the station area; n is the calculated total period.
Step 2.2: selecting the electric quantity month trend of a typical electricity stealing user, the ratio of month average current, month average voltage and month average active power of the typical electricity stealing user and the pearson correlation coefficient of the typical electricity stealing user as characteristic vectors of the typical electricity stealing user, and constructing a training sample data set D of the typical electricity stealing user according to three characteristic dimensions k ＝{P 1 ,P 2 ,P 3 ……P k }, as shown in the following table:
selecting the electric quantity month trend of the user under the abnormal platform area, the ratio of the month average voltage of the month average current of the user to the month average active power and the user pearson correlation coefficient as user characteristic vectors, and constructing a training sample data set D of the user under the abnormal platform area according to three characteristic dimensions m ＝{P 1 ,P 2 ,P 3 ……P m }, as shown in the following table:
wherein P is the identification of users under an abnormal platform area and typical electricity stealing users, m is the number of users under the abnormal platform area, k is the number of typical electricity stealing users in an anti-electricity stealing sample library, r represents the trend of electric quantity, v represents the ratio of current to voltage to active power, and ρ represents the pearson correlation coefficient of the electric quantity used and the electric quantity lost by the platform area.
Step 3: and building an improved DBSCAN clustering model, and training sample data. The method specifically comprises the following steps:
step 3.1: firstly, the user characteristic vector D under the abnormal platform area in the step 2.2 is calculated m As a DBSCAN cluster sample set, for cluster sample set D m Obtaining a clustering result through DBSCAN clustering, wherein the result comprises a clustering cluster, boundary points and noise points, and the clustering result is obtained when most users belong to normal usersBoundary points and noise points of (a) as a preliminary screening of an abnormal user set D l ；
Step 3.2: secondly, an improved DBSCAN algorithm is built, and aiming at the defect of large calculation amount caused by random selection of core points of the DBSCAN algorithm, the method is improved to select a typical electricity stealing user sample data set D k The core points used as the DBSCAN algorithm are clustered, and the specific clustering algorithm comprises the following steps:
input: sample set d=d k ∪D l A neighborhood parameter (neighborhood radius e, number of samples in neighborhood threshold mints),
(1) The typical electricity stealing user data is selected as a core object in the improved DBSCAN algorithm, namely, the core object omega=D is initialized k ；
(2) The method comprises the following steps of finding out density reachable objects with the number of samples being greater than or equal to MinPts in the omega neighborhood radius epsilon of the core object through iteration. Initializing cluster number k=0, initializing unvisited sample set Γ=d, and initializing cluster(/>Null value);
(3) Judging whether the core object set omega is null value, if soThe algorithm ends, otherwise, enter step (4);
(4) Selecting a first core object P from a core object set omega, calculating density reachable points in a neighborhood radius epsilon of the object P, initializing a set N= { P } of a current core object cluster, initializing a class sequence number j=j+1, updating a non-accessed sample set Γ=Γ - { P }, and initializing a current cluster sample set C j ＝{P}；
(5) Judging whether the set N of the current core object cluster is empty, if soThen the current cluster C k After the completion of the generation, the cluster division set c= { C is updated 1 ,C 2 ,C 3 ……C j Update core object set Ω=Ω -C j And (3) entering a step; otherwise, entering a step (6);
(6) Randomly selecting a core object P ' from a set N of current core object clusters, finding out all sub-sample sets N ' = { P ' } in an epsilon neighborhood through a sample number threshold in the neighborhood, and updating a current cluster sample set C j ＝C j Updating the unvisited sample set Γ=Γ -N ', updating the current cluster core object set n=n- (N ' Σ) -P ', and entering step (5);
and (3) ending the algorithm, and outputting a cluster C= { C of model partitioning 1 ,C 2 ,C 3 ……C j }。
Step 4: and measuring the clustering quality by adopting the contour coefficient, judging the clustering quality as an effective cluster when the contour coefficient of the clustering cluster is within a certain range, and finally judging the object in the effective cluster as a suspected electricity stealing user. The method specifically comprises the following steps:
step 4.1: the clustering quality of the DBSCAN algorithm is improved by adopting contour coefficient measurement, and the contour coefficients of all sample points in each cluster in the clustering cluster C output in the step 3.2 are calculated according to the definition of the contour coefficients, wherein the calculation steps are as follows:
(1) Computing cluster C j Average distance a between each sample point P in the cluster and other sample points in the same cluster;
(2) Computing cluster C j Average distance b between each sample point P in the cluster and all sample points in other clusters;
(3) Computing cluster C j The profile coefficient of each sample point P within the range is defined by:
wherein a is the average distance from the sample point P to other sample points in the same cluster, b is the average distance from the sample point P to the sample points in other clusters, and S is the contour coefficient of the sample point P;
(4) Calculating the average value of the contour coefficients of all sample points in a single cluster
Step 4.2: as can be seen from the definition of the profile factor, the value of the profile factor is between [ -1,1]The closer the profile coefficient value is to 1, the better the cohesion and separation are, and therefore, the average value of the profile coefficients of all points in a single cluster is usedTo measure the cluster quality of this cluster, when +.>When the cluster is judged to be an effective cluster, the effective cluster C is judged to be effective because the objects in the effective cluster comprise typical electricity stealing users and user objects under abnormal areas j Training sample data set D of user under constant area m And solving an intersection, wherein the obtained intersection result is a suspected electricity stealing user.
Example 2:
the embodiment introduces a user electricity stealing analysis device based on improved DBSCAN clustering, which comprises a processor and a storage medium.
The storage medium is for storing instructions.
The processor is operative in accordance with the instructions to perform the steps of a user electricity theft analysis method based on improved DBSCAN clustering as described above.
Example 3:
the present embodiment introduces a computer readable storage medium having stored thereon a computer program which when executed by a processor implements the steps of a user electricity theft analysis method based on improved DBSCAN clustering as described above.
Example 4:
for a more detailed description of the process according to the invention, example data are described below.
(1) Taking a district within the jurisdiction of a power supply company in a certain province as an example, selecting 25 districts of an abnormal district exceeding the theoretical line loss rate, taking 5000 users under the abnormal district and 260 typical electricity larceny users in a selected electricity larceny prevention sample library as training samples, selecting electricity larceny characteristic values according to the electricity consumption related data of the users under the abnormal district and the related data of the typical electricity larceny users, and constructing a training sample set, wherein the training sample set is shown in tables 1 and 2:
table 1 shows a typical electricity stealing user training sample set D k ：
Table 2 shows a user training sample set D for abnormal regions m ：
(2) Clustering training sample data by improving a DBSCAN algorithm, and performing quality evaluation on the clustering clusters by using contour coefficients to obtain the clustering clusters and the contour coefficients, wherein the concrete see Table 3:
table 3 is the cluster result set and the average value of the sample profile coefficients in the cluster:
(3) Obtaining a cluster C according to the average value of the contour coefficients being larger than 0.8 1 、C 2 、C 4 、C 5 、C 7 As the effective clusters, the effective cluster result set and the training sample data set D of the users under the normal platform area are obtained because the objects in the effective clusters comprise typical electricity stealing users and the users under the abnormal platform area m Intersection is calculated, and a suspected electricity stealing user list is obtained as shown in table 4:
table 4 is a list of suspected fraudulent users:
it will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is only a preferred embodiment of the invention, it being noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the invention.
Claims (8)
1. A user electricity stealing analysis method based on improved DBSCAN clustering is characterized in that: the method comprises the following steps:
step 1: acquiring lost electric quantity of an abnormal platform area, power consumption related data of users under the abnormal platform area and related data of typical power stealing users in a power stealing prevention sample library;
step 2: calculating a training sample data set of the user under the abnormal area according to the lost electric quantity of the abnormal area, the related data of the power consumption of the user under the abnormal area and the related data of the typical power stealing user in the anti-power stealing sample library;
step 3: according to the training sample data set of the user under the abnormal platform region, the training sample data set of the typical electricity stealing user obtains a cluster of the user under the abnormal platform region and the typical electricity stealing user through an improved DBSCAN cluster model;
step 4: calculating the average value of the contour coefficients of each cluster in the cluster, and acquiring an effective cluster in the cluster according to the average value of the contour coefficients;
step 5: and acquiring intersection of the users in the effective clusters and the users under the abnormal area to obtain suspected electricity stealing users under the abnormal area.
2. The user electricity theft analysis method based on improved DBSCAN clustering of claim 1, wherein: the step 1 specifically includes:
step 1.1: acquiring the line loss rate of each station area, defining the station area as an abnormal station area when the line loss rate of the station area exceeds the theoretical line loss rate, and acquiring the month loss electric quantity of the abnormal station area;
step 1.2: acquiring month electricity consumption, month trend of electric quantity, month average voltage, month average current and month average active power of a user under an abnormal platform area;
step 1.3: and acquiring the month electricity consumption, the month electricity tendency, the month average voltage, the month average current and the month average active power of a typical electricity stealing user in an electricity stealing prevention sample library.
3. The user electricity theft analysis method based on improved DBSCAN clustering of claim 2, wherein: the step 2 specifically includes:
step 2.1: obtaining correlation coefficient of month electricity consumption of users under abnormal areas and month electricity loss of abnormal areasWherein (1)>Representing the power consumption related coefficient of the users in the ith abnormal area, wherein i epsilon 1,2,3 … …, m and m are the number of the users in the abnormal area;
step 2.1: obtaining the correlation coefficient of the month electricity consumption of a typical electricity stealing user and the month lost electricity of an abnormal platform areaWherein (1)>Representing the power consumption related coefficient of the j typical electricity stealing users, wherein j epsilon 1,2,3 … …, k and k are the number of the typical electricity stealing users;
step 2.3: taking the electric quantity month trend of the user in the ith abnormal area as a characteristic r i 1 Calculating the month average voltage, month average current and month average active power of the user in the ith abnormal platform area to obtain the characteristicsWill r i 1 、/>Feature vector P for composing user under ith abnormal platform area i 1 Acquiring feature vectors of all users under the abnormal platform region to form a training sample data set D of the users under the abnormal platform region m ，/>
Step 2.4: taking the month trend of the electricity quantity of the j-th typical electricity stealing user as a characteristicCalculating month average voltage, month average current and month average active power of the j-th typical electricity stealing user to obtain characteristic +.>Will->Characteristic vector +.>Acquiring feature vectors of all typical electricity stealing users to form a training sample data set D of the typical electricity stealing users k ，/>
4. The user electricity theft analysis method based on improved DBSCAN clustering of claim 3, wherein: the step 3 specifically includes:
step 3.1: training sample number of user under abnormal platform areaFrom set D m As a DBSCAN clustering sample set, clustering the DBSCAN clustering sample set by DBSCAN to obtain a clustering result, and taking boundary points and noise points in the clustering result as an abnormal user set D of preliminary screening l ；
Step 3.2: selecting training sample data set D of typical electricity stealing users k As a core object, a core object Ω=d is initialized k ；
Step 3.3: the method comprises the following steps of searching out a density reachable object with the number of samples being more than or equal to a number threshold MinPts in a core object omega neighborhood radius epsilon through iteration; initializing the cluster number k=0 of the clusters, initializing the unvisited sample set Γ=d, d=d k ∪D l Initializing cluster clusters Is null;
step 3.4: judging whether the core object omega is null value, if soThe algorithm ends and cluster c= { C is output 1 ,C 2 ,C 3 ……C j -a }; otherwise, enter step 3.5;
step 3.5: selecting a first core object P from a core object set omega, calculating density reachable points in a neighborhood radius epsilon of the core object P, initializing a set N= { P } of a current core object cluster, initializing a class sequence number j=j+1, updating a non-accessed sample set Γ=Γ - { P }, and initializing a current cluster sample set C j ＝{P}；
Step 3.6: judging whether the set N of the current core object cluster is empty, if soThen the current cluster C k After the completion of the generation, the cluster division set c= { C is updated 1 ,C 2 ,C 3 ……C j Update core object set Ω=Ω -C j And step 3.4; otherwise, enter step 3.7;
step 3.7: randomly selecting a core object P ' from a set N of current core object clusters, finding out all sub-sample sets N ' = { P ' } in an epsilon neighborhood through a sample number threshold Mints in the neighborhood, and updating a current cluster sample set C j ＝C j And updating the unvisited sample set Γ=Γ -N ', updating the current cluster core object set N=N- (N '. Cndot. OMEGA) -P ', and entering step 3.6.
5. The user electricity theft analysis method based on improved DBSCAN clustering of claim 4, wherein: the step 4 specifically includes:
step 4.1: calculating the average distance a between each sample point P in each cluster in the cluster C and other sample points in the same cluster;
step 4.2: calculating the average distance b between each sample point P in each cluster in the cluster C and all sample points in other clusters;
step 4.3: the contour coefficient S of each sample point P in each cluster in the cluster C is calculated, and the calculation formula of the contour coefficient is as follows:
step 4.4: calculating average value of contour coefficients of all sample points in each cluster in cluster C
Step 4.5: when (when)When the threshold delta is larger than the set threshold delta, the corresponding cluster in the cluster clusters is an effective cluster.
6. The user electricity theft analysis method based on improved DBSCAN clustering of claim 5, wherein: the delta is set to 0.8.
7. A user electricity stealing analysis device based on improved DBSCAN clustering is characterized in that: comprises a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is operative to perform the steps of the user electricity theft analysis method based on improved DBSCAN clustering as claimed in any one of claims 1-6.
8. A computer-readable storage medium, characterized by: a computer program stored thereon, which when executed by a processor, implements the steps of the user electricity theft analysis method based on improved DBSCAN clustering as claimed in any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310924145.XA CN116910596B (en) | 2023-07-26 | 2023-07-26 | User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310924145.XA CN116910596B (en) | 2023-07-26 | 2023-07-26 | User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116910596A true CN116910596A (en) | 2023-10-20 |
| CN116910596B CN116910596B (en) | 2024-06-21 |
Family
ID=88364540
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310924145.XA Active CN116910596B (en) | 2023-07-26 | 2023-07-26 | User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116910596B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110969539A (en) * | 2019-11-28 | 2020-04-07 | 温岭市非普电气有限公司 | Photovoltaic electricity stealing discovery method and system based on curve morphological analysis |
| CN111539840A (en) * | 2019-12-04 | 2020-08-14 | 国网天津市电力公司电力科学研究院 | Electricity stealing detection method and system fusing clustering and density estimation |
| WO2020191663A1 (en) * | 2019-03-27 | 2020-10-01 | 华北电力大学扬中智能电气研究中心 | Method and apparatus for detecting electricity consumption behavior, electronic device and storage medium |
| WO2020252785A1 (en) * | 2019-06-21 | 2020-12-24 | 西门子股份公司 | Abnormal electricity use recognition method and device, and computer readable storage medium |
| WO2022110557A1 (en) * | 2020-11-25 | 2022-06-02 | 国网湖南省电力有限公司 | Method and device for diagnosing user-transformer relationship anomaly in transformer area |
| WO2023109527A1 (en) * | 2021-12-17 | 2023-06-22 | 广东电网有限责任公司东莞供电局 | Electricity theft behavior detection method and apparatus, computer device and storage medium |
-
2023
- 2023-07-26 CN CN202310924145.XA patent/CN116910596B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020191663A1 (en) * | 2019-03-27 | 2020-10-01 | 华北电力大学扬中智能电气研究中心 | Method and apparatus for detecting electricity consumption behavior, electronic device and storage medium |
| WO2020252785A1 (en) * | 2019-06-21 | 2020-12-24 | 西门子股份公司 | Abnormal electricity use recognition method and device, and computer readable storage medium |
| CN110969539A (en) * | 2019-11-28 | 2020-04-07 | 温岭市非普电气有限公司 | Photovoltaic electricity stealing discovery method and system based on curve morphological analysis |
| CN111539840A (en) * | 2019-12-04 | 2020-08-14 | 国网天津市电力公司电力科学研究院 | Electricity stealing detection method and system fusing clustering and density estimation |
| WO2022110557A1 (en) * | 2020-11-25 | 2022-06-02 | 国网湖南省电力有限公司 | Method and device for diagnosing user-transformer relationship anomaly in transformer area |
| WO2023109527A1 (en) * | 2021-12-17 | 2023-06-22 | 广东电网有限责任公司东莞供电局 | Electricity theft behavior detection method and apparatus, computer device and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 李顺杰 等: "基于分相线损的低压配电台区反窃电分析", 河北电力技术, vol. 42, no. 3, 30 June 2023 (2023-06-30), pages 75 - 78 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116910596B (en) | 2024-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105512799B (en) | Power system transient stability evaluation method based on mass online historical data | |
| Wu et al. | A data mining approach for spatial modeling in small area load forecast | |
| CN110288039B (en) | Electricity stealing detection method based on user electricity load characteristics | |
| CN105005825B (en) | A kind of method and system based on dynamic neural network prediction photovoltaic power | |
| CN108198408B (en) | Self-adaptive anti-electricity-stealing monitoring method and system based on electricity information acquisition system | |
| CN103177188A (en) | Electric system load dynamic characteristic classifying method based on characteristic mapping | |
| CN110119556B (en) | Space-time evolution analysis method for regional water source conservation function | |
| CN110569883B (en) | Air quality index prediction method based on Kohonen network clustering and Relieff feature selection | |
| CN114511181A (en) | Water pollution environment-friendly calibration method and device based on power grid and tax data fusion | |
| CN112330077A (en) | Power load prediction method, power load prediction device, computer equipment and storage medium | |
| CN110751176A (en) | Lake water quality prediction method based on decision tree algorithm | |
| CN115603320A (en) | Real-time sensing and online evaluation method and system based on wide area measurement data | |
| CN114266416A (en) | Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium | |
| CN115563477A (en) | Harmonic data identification method and device, computer equipment and storage medium | |
| CN111539840A (en) | Electricity stealing detection method and system fusing clustering and density estimation | |
| CN111062539A (en) | Total electric quantity prediction method based on secondary electric quantity characteristic clustering analysis | |
| CN112528762B (en) | Harmonic source identification method based on data correlation analysis | |
| CN109858667A (en) | It is a kind of based on thunder and lightning weather to the short term clustering method of loading effects | |
| CN116910596B (en) | User electricity stealing analysis method, device and storage medium based on improved DBSCAN clustering | |
| CN112819208A (en) | Spatial similarity geological disaster prediction method based on feature subset coupling model | |
| CN115166625A (en) | Intelligent ammeter error estimation method and device | |
| Wang et al. | A Novel Multi‐Input AlexNet Prediction Model for Oil and Gas Production | |
| CN106816871B (en) | State similarity analysis method for power system | |
| CN111985526B (en) | Similar scene clustering-based trailing interval management strategy generation method and system | |
| CN109871998B (en) | Power distribution network line loss rate prediction method and device based on expert sample library |
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 |