CN112884000A - Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining - Google Patents
Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining Download PDFInfo
- Publication number
- CN112884000A CN112884000A CN202110056298.8A CN202110056298A CN112884000A CN 112884000 A CN112884000 A CN 112884000A CN 202110056298 A CN202110056298 A CN 202110056298A CN 112884000 A CN112884000 A CN 112884000A
- Authority
- CN
- China
- Prior art keywords
- user
- load
- data
- electricity
- power
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000003745 diagnosis Methods 0.000 title claims abstract description 29
- 238000007689 inspection Methods 0.000 title claims abstract description 28
- 238000007418 data mining Methods 0.000 title claims abstract description 20
- 230000005611 electricity Effects 0.000 claims abstract description 92
- 230000002159 abnormal effect Effects 0.000 claims abstract description 51
- 238000013528 artificial neural network Methods 0.000 claims abstract description 23
- 238000011156 evaluation Methods 0.000 claims abstract description 11
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000012216 screening Methods 0.000 claims abstract description 5
- 230000006399 behavior Effects 0.000 claims description 31
- 238000001514 detection method Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 9
- 230000010354 integration Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 230000001419 dependent effect Effects 0.000 claims description 4
- 238000009472 formulation Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000007667 floating Methods 0.000 claims description 3
- 239000013589 supplement Substances 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 abstract description 15
- 238000011835 investigation Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 238000007621 cluster analysis Methods 0.000 abstract 1
- 238000011161 development Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/283—Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Evolutionary Biology (AREA)
- Automation & Control Theory (AREA)
- Computational Mathematics (AREA)
- Fuzzy Systems (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A power utilization inspection intelligent diagnosis method and a diagnosis system based on data mining are disclosed, wherein the diagnosis method comprises the following steps: 1. collecting power utilization information of a user; 2. preprocessing each data source acquired in the step 1; 3. performing cluster analysis on the electricity consumption information data preprocessed in the step 2 by adopting an FCM clustering algorithm to obtain a typical load characteristic curve; 4. calculating the similarity between the user and the typical load characteristic curve and screening abnormal electricity consumption suspicion users; 5. constructing an abnormal electricity utilization evaluation index system; 6. constructing an anti-electricity-stealing behavior recognition model by using a fuzzy neural network; 7. and (6) predicting the electricity stealing behavior by using the electricity stealing behavior identification model obtained in the step 6. The diagnosis system comprises a data layer, an application layer and a display layer. The method and the system can quickly determine the functions of users with abnormal electricity stealing, enhance the security of abnormal electricity utilization activities at the investigation department, realize the concealed electricity investigation and greatly improve the efficiency of electricity stealing at the low voltage investigation department.
Description
Technical Field
The invention relates to the technical field of electric power safety, in particular to an intelligent diagnosis method and a diagnosis system for power utilization inspection based on data mining.
Background
The traditional electricity charge collection mode is manual meter reading, physical labor is strong, the number of foreign-involved personnel is insufficient, the problem of electricity stealing inspection is also borne by the foreign-involved personnel, and the number of electricity utilization inspection personnel is difficult to meet the requirement; secondly, because the thought concept of traditional power consumption management is deeply rooted, power supply enterprises pay little attention to the behavior of stealing electricity, cause the power consumption to check and strike the behavior of stealing electricity and have done good job with ease, has helped the behavior of stealing electricity. With social progress and economic development, the quantity of power consumption of users is continuously increased, the demand of social production and life on electric energy is increasingly large, electric power enterprises introduce intelligent electric meters, and remote cost control is effectively realized through a power consumption information acquisition system and a marketing service application system, however, the problem of electricity stealing investigation is not solved, and under the drive of economic benefits, electricity stealers are not limited to past residents, individuals and the like, are gradually developed to collective enterprises, China and foreign cooperation enterprises and the like, the development speed is very high, and the power supply safety and the order of the electric power enterprises are seriously disturbed. In addition, along with scientific and technological change day by day, the high-tech means is stolen electric personnel and is used widely, along with the intellectuality, the technological development of stealing electric technology for high-tech content steals electric mode more and more, like wireless remote control, wired remote control etc. this kind steals electric means and often hides very much, and traditional power consumption inspection method can't detect at all, and the comprehensive quality of power consumption inspection personnel is lower, is difficult to satisfy the modernization demand of power consumption inspection and anti-electricity-stealing work.
In the process of rapid development of the power industry, many power technologies have effectively advanced and perfected, but the development of power detection and anti-theft technologies and equipment is relatively delayed. At present, the electricity utilization detecting instruments and equipment provided by power grid enterprises in China can only detect the existing and relatively common electricity stealing means, but can hardly detect newly developed and emerging electricity stealing means.
Therefore, an effective method is urgently needed to be adopted, data provided by the existing system of the power enterprise are used for realizing anti-electricity-stealing analysis, probability speculation and diagnosis are carried out on the behavior of the electricity-stealing suspected user, the major electricity-stealing suspected user is accurately identified, the work effect of anti-electricity-stealing work is improved, and the high-efficiency supervision of the power enterprise in China on electric energy output is enhanced. By adopting a powerful electricity stealing monitoring and identifying means, the punishment and punishment force of electricity stealing is increased, the normal power supply and utilization order is maintained, and the business benefits of a company are guaranteed.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide the power utilization inspection intelligent diagnosis system based on data mining, which utilizes massive power utilization information and carries out mining analysis on data based on a scientific means, thereby greatly improving the efficiency of anti-theft and power utilization inspection.
The invention adopts the following technical scheme:
a power utilization inspection intelligent diagnosis method based on data mining comprises the following steps:
step 1: collecting power utilization information of each user; the electricity utilization information comprises instantaneous electricity quantity and electricity utilization load of a user;
step 2: preprocessing the electricity utilization information acquired in the step 1, including performing data synchronization on the electricity utilization information and completing the supplement of missing data;
and step 3: clustering and analyzing the electricity utilization information data preprocessed in the step 2 by using an FCM clustering algorithm to obtain clustering center curves of various typical loads, wherein each curve represents a typical load characteristic curve and corresponds to the electricity utilization characteristics of each type of user;
and 4, step 4: calculating the similarity between the load characteristic curve of the user to be detected and the typical load characteristic curve obtained in the step 3, judging whether the user has suspicion of abnormal electricity utilization or not through similarity matching, and screening out the suspected users of abnormal electricity utilization;
and 5: calculating the abnormal electricity suspicion user screened in the step 4 to obtain characteristic index data of the user, and constructing an abnormal electricity evaluation index system;
the characteristic index data comprises an electric quantity index, a load index, a line loss index, an alarm index, a contract capacity ratio, a voltage three-phase unbalance rate and a power factor;
step 6: constructing an anti-electricity-stealing behavior recognition model by using a fuzzy neural network;
and 7: and (6) predicting the electricity stealing behavior by using the electricity stealing behavior identification model obtained in the step 6.
The intelligent field detection equipment in the step 1 comprises an intelligent detection terminal palm machine and an intelligent current clamp.
The data source preprocessing in the step 2 comprises the following steps:
step 201: storing the collected electricity utilization information in the step 1 to a data loading source, performing data synchronization by using data of a data integration data loading source of the DataWorks, and completing the operations of data condition extraction increment synchronization, loading and conversion in the synchronization process;
step 202: processing missing values of users with missing power utilization data, judging the users with missing data for more than 2 hours as abnormal users, and interpolating by adopting a Lagrange interpolation method when the users with missing data for less than or equal to 2 hours are missing;
step 203: the data are normalized, and the processing formula is as follows:
wherein L (m, y) represents a normalized value of the electric load, L' (m, y) represents an actual value of the electric load, and L (m, y)maxRepresenting the maximum value in the sequence of actual electrical load values, L (m, y)minRepresenting the minimum value in the sequence of actual electrical load values.
Data loading sources include MySQL, Oracle, Kafka, DataHub, LogHub, and PolarDB;
the data warehouse includes MaxCommute, Hologres, Kafka, and DataHub.
The FCM clustering algorithm in the step 3 divides the users into five categories; the first type: the accumulated power load of a user in 80% of 24 hours of the day is not higher than 60% of the average value of the daily power loads of all users to be diagnosed, the slope of a load time characteristic curve of the user in the day is smaller than 1, and the hourly power load of the user in the evening period is in an hourly ascending trend; the second type: the daily power load of the user is not lower than 80% of the average value of the daily power loads of all users to be diagnosed, and the average value of the absolute values of the power load difference values of the users in adjacent hours at night is lower than 20% of the daily power load of the users; in the third category: the user has two peak values in the morning and at night when the electricity load is used every hour; the fourth type: the daily electric load of the user is between 90 percent and 100 percent of the average value of the daily electric loads of all users to be diagnosed, and the slope of a load time characteristic curve of the current day of the user is less than 0.5 in 90 percent of the time; the fifth type: the daily power load of the user is between 60% and 90% of the average value of the daily power loads of all users to be diagnosed, and the slope of a load time characteristic curve of the user on the same day is more than 1 in 70% of the time;
"morning" means from 6 am to 12 am, "evening" means from 6 pm to 12 pm of the day;
the independent variable of the characteristic curve is time, the unit is hour, and the dependent variable is electric load.
The similarity in step 4 is calculated by measuring the distance between two points in the n-dimensional space:
for the similarity matching, the Euclidean distance dist (x, y) is calculated and the reciprocal is calculated
The matching degree ma is controlled to be 0,1]To (c) to (d); the larger the ma value is, the greater the similarity between the power load characteristic curve of the user and the typical load characteristic curve is; when the matching degree ma values of the user and the five typical load characteristic curves are all smaller than 0.5, the user is judged to be a suspected user, otherwise, the user is a normal power utilization user; the five typical load characteristic curves are curves representing the electricity utilization characteristics of the five types of users obtained by the FCM clustering algorithm in the step 3.
The electric quantity index in the step 5 is the average electric quantity used for n days by the user;
the load index is a load average moving value which is a moving average value of each moment n-1 days before the real-time load is calculated, and the load average moving value formula is as follows:
wherein L isijIs the load at time j on day i,
is the average load at time j from the previous n days to t days;
and then calculating the load gradient, wherein the formula is as follows:
wherein max (avg) is the maximum load average moving value in t consecutive days;
the alarm index is that when the voltage or current phase and polarity indexes are abnormal, the acquisition terminal sends out an alarm signal, and the alarm frequency of the terminal is taken as the index;
the contract capacity ratio is the total capacity of the user power receiving equipment permitted by the power supply department and in the power supply and utilization contract, and if the total capacity exceeds a specified range, the user has the risk of abnormal power utilization.
The fuzzy neural network in the step 6 has 5 layers in total, and the training steps are as follows:
step 601: inputting m indexes in the abnormal electricity utilization evaluation index system obtained in the step 5;
step 602: fuzzifying input variables, wherein each input index corresponds to 3 fuzziness degrees, the total number of the fuzziness degrees is 3m, and the activation function of each node is a membership function of the fuzziness degree of each input variable x
Wherein, aijAs the central value of the membership function, bijInitializing the center of the membership function and the width of the membership function in the fuzzy neural network by a random function as the width of the membership function; the output after fuzzification is [0, 1]]A value in between;
step 603: performing 'and' operation of fuzzy inference, wherein the step is completed in a fuzzy inference layer of the fuzzy neural network, and the node number of the fuzzy inference layer is 3 in totalmI.e. have 3mA rule; theThe output value of the step is
Step 604: defuzzification, normalization calculation, the output of the layer is expressed as
P is the value of the ambiguity;
step 605: converting the output of each node of the deblurring layer into one in an output layer, namely a suspicion coefficient of abnormal electricity utilization behavior, and adopting a weighted linear summation method to obtain an output value of
Wherein ω isjkThe connection weight between the de-fuzzy layer and the output layer is set as a random floating point number in the (0, 1) interval, and the sum is 1.
The maximum error of a single sample of the fuzzy neural network is set to 0.001, the maximum error of the system is set to 0.1, and the neural network iterates 2000 times.
The abnormal electricity consumption suspicion coefficient is the suspicion degree of whether the user has abnormal electricity consumption behaviors or not, specifically, the suspicion coefficient is the suspicion coefficient when the quantitative index is reached, the suspicion coefficient value range is [0, 1], and the size represents the possibility of the user having the abnormal electricity consumption behaviors; wherein, when the value of the suspicion coefficient is in the range of [0,0.5), the suspicion coefficient belongs to normal power utilization; when the value of the suspicion coefficient is in the range of [0.5,1], the power utilization is abnormal.
The data layer comprises a data acquisition module and a data preprocessing module;
the application layer comprises an anti-electricity-stealing behavior identification model based on a fuzzy neural network, a user electricity consumption behavior comprehensive analysis module, a statistical analysis and report formulation module and an auxiliary information query module;
the display layer comprises a system management page, a data integration and query page and a power utilization abnormity analysis page.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention reduces the consumption of manpower and material resources in the electricity utilization inspection and electricity stealing prevention work of the power company, saves time and improves the accuracy of judging electricity stealing behaviors. The high-efficiency interaction of technologies such as data mining and artificial intelligence is realized, and the fusion application of various services in the power industry is promoted.
2. According to the invention, through a friendly man-machine interaction page, the power utilization condition of the user, the intelligent identification result of the power stealing behavior and the like are clearly fed back to the power utilization detection personnel in real time, the burden of the power utilization detection personnel is reduced, the work of preventing power stealing is more efficient, intelligent and standard, various power stealing behaviors are effectively restrained, favorable evidence is provided for compensating the loss of electric quantity and penalty payment, and loss reduction gain is provided for power companies.
Drawings
FIG. 1 is a system architecture diagram of the present invention;
FIG. 2 is a process flow diagram of feature calculation and feature selection in accordance with the present invention;
FIG. 3 is a flow chart of the present invention for diagnosing the behavior of electricity stealing prevention based on fuzzy neural network;
FIG. 4 is a general flow diagram of the intelligent diagnostic method of the present invention.
Detailed Description
The present application is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present application is not limited thereby.
As shown in fig. 1 to 3, the present invention employs the following embodiments:
an intelligent diagnosis method for power utilization inspection based on data mining comprises the following steps:
step 1, acquiring power utilization information through on-site detection intelligent equipment and a metering automation system, and acquiring user information from a marketing system; the user information comprises personal information such as user age, area and industry, and is used as a supplement for subsequent user electricity utilization comprehensive analysis, statistical tabulation and auxiliary information query.
The intelligent field detection equipment comprises an intelligent detection terminal palm machine and an intelligent current clamp, and the electricity utilization information comprises instantaneous electric quantity and electricity utilization load of a user;
step 2, preprocessing the collected data in the step 1, including data integration, power consumption data missing value processing, power consumption data noise value processing and normalization processing, and specifically including the following steps:
step 201: carrying out data synchronization on the electricity utilization information acquired in the step 1 by using a data set of DataWorks, and completing incremental synchronization, loading and conversion work on the data in the synchronization process; preprocessing the synchronous result, mainly completing type conversion and filling missing values, and loading the preprocessed result into a data warehouse for persistent storage;
data loading sources include MySQL, Oracle, Kafka, DataHub, LogHub, and PolarDB;
the data warehouse includes: MaxCommute, Hologres, Kafka, and DataHub;
the data conversion comprises data filtering, character string replacement and the like.
Step 201: and (4) carrying out missing value processing on users with missing power utilization data, judging the users with missing data for more than 2 hours as abnormal users, and carrying out interpolation by adopting a Lagrangian interpolation method when the users with missing data for less than or equal to 2 hours are missing.
Step 202: processing abnormal points of data with a time interval of two hours, correcting the data according to a missing value processing method, and not clearing the abnormal points of other types;
step 203: the data are normalized, and the processing formula is as follows:
wherein L (m, y) represents a normalized value of the electric load, L' (m, y) represents an actual value of the electric load, and L (m, y)maxRepresenting the maximum value in the sequence of actual electrical load values, L (m, y)minRepresenting the minimum value in the sequence of actual electrical load values.
Step 3, performing feature calculation on the structured data in the MaxCommute table by using an FCM clustering algorithm, and storing the result in the corresponding MaxCommute table; and obtaining representative clustering center curves, wherein each curve represents a type of typical load characteristic curve and corresponds to the electricity utilization characteristics of each type of users. The independent variable of the characteristic curve is time, and the dependent variable is daily power consumption load.
The FCM clustering algorithm in the step 3 divides the users into five categories; the first type: the accumulated power load of a user in 80% of 24 hours of the day is not higher than 60% of the average value of the daily power loads of all users to be diagnosed, the slope of a load time characteristic curve of the user in the day is smaller than 1, and the hourly power load of the user in the evening period is in an hourly ascending trend; the second type: the daily power load of the user is not lower than 80% of the average value of the daily power loads of all users to be diagnosed, and the average value of the absolute values of the power load difference values of the users in adjacent hours at night is lower than 20% of the daily power load of the users; in the third category: the user has two peak values in the morning and at night when the electricity load is used every hour; the fourth type: the daily electric load of the user is between 90 percent and 100 percent of the average value of the daily electric loads of all users to be diagnosed, and the slope of a load time characteristic curve of the current day of the user is less than 0.5 in 90 percent of the time; the fifth type: the daily power load of the user is between 60% and 90% of the average value of the daily power loads of all users to be diagnosed, and the slope of a load time characteristic curve of the user on the same day is more than 1 in 70% of the time;
"morning" means from 6 am to 12 am, "evening" means from 6 pm to 12 pm of the day;
the independent variable of the characteristic curve is time, the unit is hour, and the dependent variable is electric load.
And 4, step 4: calculating the similarity between the load characteristic curve of the user to be detected and the typical load characteristic curve obtained in the step 3, judging whether the user has suspicion of abnormal electricity utilization or not through similarity matching, and screening out the suspected users of abnormal electricity utilization;
the similarity is obtained by calculating a distance formula between two points in the n-dimensional space:
by calculating the Euclidean distance dist (x, y) and inverting it
The matching degree ma is controlled to be 0,1]In the meantime. The larger the ma value is, the greater the similarity between the power load characteristic curve of the user and the typical load characteristic curve is; when the matching degree ma values of the user and the five typical load characteristic curves are all smaller than 0.5, the user is judged to be a suspected user, otherwise, the user is a normal power utilization user; the five typical load characteristic curves are curves representing the electricity utilization characteristics of the five types of users obtained by the FCM clustering algorithm in the step 3.
And 5, constructing an abnormal power utilization evaluation index system by performing data reconstruction on the collected characteristic index data to form a set of complete abnormal power utilization evaluation index system:
eight characteristic indexes, namely an electric quantity index, a load index, a line loss index, an alarm index, a contract capacity ratio, a voltage three-phase unbalance rate and a power factor of a user are collected, and an abnormal power utilization evaluation index system is constructed.
Electric quantity index: calculating the average electricity consumption of n days of the user
Load index: calculating the moving average value of each moment n-1 days before the real-time load, wherein the formula is as follows:
wherein L isijIs the load at time j on day i,
is the average load at time j from the first n days to t days.
And then calculating the load gradient, wherein the formula is as follows:
where max (avg) is the maximum load average moving value for t consecutive days.
Alarm indexes are as follows: when the indexes such as voltage or current phase, polarity and the like are abnormal, the acquisition terminal can send out a warning signal, and the number of terminal alarms is used as the index.
Contract capacity ratio: if the total capacity of the user power receiving equipment permitted by the power supply department and in the power supply and utilization contract exceeds the specified range, the user has the risk of abnormal power utilization
Step 6, dividing the characteristic index data, namely sample data into two parts, taking 90% of the data as training data for training a network, and taking 10% of the data as a test sample for testing and verifying the trained network model;
a prediction model is constructed by utilizing a fuzzy neural network, the input value of the fuzzy neural network is eight characteristic indexes in an abnormal electricity utilization evaluation index system, the output value is an abnormal electricity utilization suspicion coefficient, the meaning of the abnormal electricity utilization suspicion coefficient is that whether the user has abnormal electricity utilization behavior, namely, the suspicion degree, specifically, the output value is the suspicion coefficient when the quantitative index is the suspicion coefficient, the dereferencing range of the suspicion coefficient is [0, 1], and the size of the suspicion coefficient represents the possibility of the abnormal electricity utilization behavior of the user.
When the value of the suspicion coefficient is in the range of [0,0.5), the suspicion coefficient belongs to normal power utilization;
when the value of the suspicion coefficient is in the range of [0.5,1], the suspicion coefficient belongs to abnormal power utilization;
the fuzzy neural network in the step 6 has 5 layers in total, and the training steps are as follows:
step 601: inputting m indexes in the abnormal electricity utilization evaluation index system obtained in the step 5;
step 602: fuzzifying input variables, wherein each input index corresponds to 3 fuzziness degrees, the total number of the fuzziness degrees is 3m, and the activation function of each node is a membership function of the fuzziness degree of each input variable x
Wherein, aijAs the central value of the membership function, bijFor the width of the membership function, by pairs of random functionsInitializing the center of a membership function and the width of the membership function in the fuzzy neural network; the output after fuzzification is [0, 1]]A value in between;
step 603: performing 'and' operation of fuzzy inference, wherein the step is completed in a fuzzy inference layer of the fuzzy neural network, and the node number of the fuzzy inference layer is 3 in totalmI.e. have 3mA rule; the output value of this step is
Step 604: defuzzification is carried out, normalization calculation is achieved, and oscillation caused by overlarge correction quantity in the learning process is avoided. The output of this layer can be represented as
P is the value of the ambiguity, P in the method being 3;
step 605: converting the output of each node of the deblurring layer into one in an output layer, namely a suspicion coefficient of abnormal electricity utilization behavior, and adopting a weighted linear summation method to obtain an output value of
Wherein ω isjkThe connection weight between the de-fuzzy layer and the output layer is set as a random floating point number in a (0, 1) interval, the sum is 1, and k is 1;
where the single sample maximum error is set to 0.001, the system maximum error is set to 0.1, and the neural network iterates 2000 times.
And 7: and (6) predicting the electricity stealing behavior by using the electricity stealing behavior identification model obtained in the step 6.
And comparing the test result with the expected result, and analyzing the diagnosis effect of the model through the comparison between the test result and the expected result, the output error rate and the misjudgment rate.
The invention also comprises a diagnosis system based on the intelligent power utilization inspection obtained by data mining, which comprises a data layer, an application layer and a display layer;
the data layer comprises a data acquisition module and a data preprocessing module;
the application layer comprises an anti-electricity-stealing behavior identification model based on a fuzzy neural network, a statistical analysis and report formulation module and an auxiliary information query module;
the statistical analysis and report formulation module integrates and tabulates the information of the five typical user numbers obtained by FCM clustering, the proportion of the five typical user numbers to the total user, the evaluation index data of each abnormal power consumption of the user, whether the abnormal power consumption exists and the like.
The auxiliary information inquiry is convenient for the detection personnel to inquire the required power utilization information and the user information, and aims to make full use of data.
The display layer comprises a system management page, a data integration and query page and a power utilization abnormity analysis page.
Can develop a set of application system platform based on data mining under anti-electricity-stealing scene through the implementation of project, regard the electricity consumption inspection as the research scene, carry on the aircraft carrier of big data, adopt the thing networking, big data analysis, leading edge technologies such as artificial intelligence degree of depth study, innovate traditional power consumption behavior monitoring means, research and development intelligence anti-electricity-stealing complete set of system that detects, realize the novel data acquisition to the user, read in batches, check, intelligent diagnosis analysis, screening, reduce and confirm to look for the scope of stealing, confirm the unusual user's of electricity-stealing function fast, strengthen looking for an abnormal situation of electricity consumption action confidentiality, realize that the disguised is looked for and is stolen, greatly improve the efficiency that the electricity was stolen to low pressure looking for a situation. The method provides favorable evidence for supplementing lost electric quantity and penalty, improves economic and social benefits, and promotes the safe and stable development of electric power enterprises.
The present applicant has described and illustrated embodiments of the present invention in detail with reference to the accompanying drawings, but it should be understood by those skilled in the art that the above embodiments are merely preferred embodiments of the present invention, and the detailed description is only for the purpose of helping the reader to better understand the spirit of the present invention, and not for limiting the scope of the present invention, and on the contrary, any improvement or modification made based on the spirit of the present invention should fall within the scope of the present invention.
Claims (11)
1. The intelligent power utilization inspection diagnosis method based on data mining is characterized by comprising the following steps of:
step 1: collecting power utilization information of each user; the electricity utilization information comprises instantaneous electricity quantity and electricity utilization load of a user;
step 2: preprocessing the electricity utilization information acquired in the step 1, including performing data synchronization on the electricity utilization information and completing the supplement of missing data;
and step 3: clustering and analyzing the electricity utilization information data preprocessed in the step 2 by using an FCM clustering algorithm to obtain clustering center curves of various typical loads, wherein each curve represents a typical load characteristic curve and corresponds to the electricity utilization characteristics of each type of user;
and 4, step 4: calculating the similarity between the load characteristic curve of the user to be detected and the typical load characteristic curve obtained in the step 3, judging whether the user has suspicion of abnormal electricity utilization or not through similarity matching, and screening out the suspected users of abnormal electricity utilization;
and 5: calculating the abnormal electricity suspicion user screened in the step 4 to obtain characteristic index data of the user, and constructing an abnormal electricity evaluation index system;
the characteristic index data comprises an electric quantity index, a load index, a line loss index, an alarm index, a contract capacity ratio, a voltage three-phase unbalance rate and a power factor;
step 6: constructing an anti-electricity-stealing behavior recognition model by using a fuzzy neural network;
and 7: and (6) predicting the electricity stealing behavior by using the electricity stealing behavior identification model obtained in the step 6.
2. The intelligent power inspection diagnosis method based on data mining as claimed in claim 1, wherein:
the intelligent field detection equipment in the step 1 comprises an intelligent detection terminal palm machine and an intelligent current clamp.
3. The intelligent power inspection diagnosis method based on data mining as claimed in claim 1, wherein:
the data source preprocessing in the step 2 comprises the following steps:
step 201: storing the collected electricity utilization information in the step 1 to a data loading source, performing data synchronization by using data of a data integration data loading source of the DataWorks, and completing the operations of data condition extraction increment synchronization, loading and conversion in the synchronization process;
step 202: processing missing values of users with missing power utilization data, judging the users with missing data for more than 2 hours as abnormal users, and interpolating by adopting a Lagrange interpolation method when the users with missing data for less than or equal to 2 hours are missing;
step 203: the data are normalized, and the processing formula is as follows:
wherein L (m, y) represents a normalized value of the electric load, L' (m, y) represents an actual value of the electric load, and L (m, y)maxRepresenting the maximum value in the sequence of actual electrical load values, L (m, y)minRepresenting the minimum value in the sequence of actual electrical load values.
4. The intelligent power inspection diagnosis method based on data mining as claimed in claim 3, wherein:
the data loading source comprises MySQL, Oracle, Kafka, DataHub, LogHub and PolarDB;
the data warehouse includes MaxCommute, Hologres, Kafka, and DataHub.
5. The intelligent power inspection diagnosis method based on data mining as claimed in claim 1, wherein:
the FCM clustering algorithm in the step 3 divides the users into five classes; the first type: the accumulated power load of a user in 80% of 24 hours of the day is not higher than 60% of the average value of the daily power loads of all users to be diagnosed, the slope of a load time characteristic curve of the user in the day is smaller than 1, and the hourly power load of the user in the evening period is in an hourly ascending trend; the second type: the daily power load of the user is not lower than 80% of the average value of the daily power loads of all users to be diagnosed, and the average value of the absolute values of the power load difference values of the users in adjacent hours at night is lower than 20% of the daily power load of the users; in the third category: the user has two peak values in the morning and at night when the electricity load is used every hour; the fourth type: the daily electric load of the user is between 90 percent and 100 percent of the average value of the daily electric loads of all users to be diagnosed, and the slope of a load time characteristic curve of the current day of the user is less than 0.5 in 90 percent of the time; the fifth type: the daily power load of the user is between 60% and 90% of the average value of the daily power loads of all users to be diagnosed, and the slope of a load time characteristic curve of the user on the same day is more than 1 in 70% of the time;
the "morning" refers to from 6 am to 12 am, and the "evening" refers to from 6 pm to 12 pm of the day;
the independent variable of the characteristic curve is time, the unit is hour, and the dependent variable is electric load.
6. The intelligent power inspection diagnosis method based on data mining as claimed in claim 5, wherein:
the similarity in step 4 is calculated by measuring the distance between two points in the n-dimensional space:
for the similarity matching, the Euclidean distance dist (x, y) is calculated and the reciprocal is calculated
The matching degree ma is controlled to be 0,1]To (c) to (d); the larger the ma value is, the greater the similarity between the power load characteristic curve of the user and the typical load characteristic curve is; when the user isWhen the matching degree ma values of the five typical load characteristic curves are less than 0.5, the user is judged to be a suspected user, otherwise, the user is a normal power utilization user; the five typical load characteristic curves are curves representing the electricity utilization characteristics of the five types of users obtained by the FCM clustering algorithm in the step 3.
7. The intelligent power inspection diagnosis method based on data mining as claimed in claim 1, wherein:
the electric quantity index in the step 5 is the average electric quantity used for n days by the user;
the load index is a load average moving value which is a moving average value of each moment n-1 days before the real-time load is calculated, and the load average moving value formula is as follows:
wherein L isijIs the load at time j on day i,
is the average load at time j from the previous n days to t days;
and then calculating the load gradient, wherein the formula is as follows:
wherein max (avg) is the maximum load average moving value in t consecutive days;
the alarm index is that when the voltage or current phase and polarity indexes are abnormal, the acquisition terminal sends out an alarm signal, and the alarm frequency of the terminal is taken as the index;
the contract capacity ratio is the total capacity of the user power receiving equipment permitted by the power supply department and in the power supply and utilization contract, and if the total capacity exceeds a specified range, the user has the risk of abnormal power utilization.
8. The intelligent power inspection diagnosis method based on data mining as claimed in claim 7, wherein:
the fuzzy neural network in the step 6 has 5 layers in total, and the training steps are as follows:
step 601: inputting m indexes in the abnormal electricity utilization evaluation index system obtained in the step 5;
step 602: fuzzifying input variables, wherein each input index corresponds to 3 fuzziness degrees, the total number of the fuzziness degrees is 3m, and the activation function of each node is a membership function of the fuzziness degree of each input variable x:
wherein, aijAs the central value of the membership function, bijInitializing the center of the membership function and the width of the membership function in the fuzzy neural network by a random function as the width of the membership function; the output after fuzzification is [0, 1]]A value in between;
step 603: performing 'and' operation of fuzzy inference, wherein the step is completed in a fuzzy inference layer of the fuzzy neural network, and the node number of the fuzzy inference layer is 3 in totalmI.e. have 3mA rule; the output value of this step is
Step 604: defuzzification, normalization calculation, the output of the layer is expressed as
P is the value of the ambiguity;
step 605: converting the output of each node of the deblurring layer into one in an output layer, namely a suspicion coefficient of abnormal electricity utilization behavior, and adopting a weighted linear summation method to obtain an output value of
Wherein ω isjkThe connection weight between the de-fuzzy layer and the output layer is set as a random floating point number in the (0, 1) interval, and the sum is 1.
9. The intelligent power inspection diagnosis method based on data mining as claimed in claim 8, wherein:
the maximum error of a single sample of the fuzzy neural network is set to be 0.001, the maximum error of a system is set to be 0.1, and the neural network iterates for 2000 times.
10. The intelligent power inspection diagnosis method based on data mining as claimed in claim 9, wherein:
the abnormal electricity consumption suspicion coefficient is the suspicion degree of whether the user has abnormal electricity consumption behaviors or not, specifically, the suspicion coefficient is the suspicion coefficient when the quantitative index is reached, the suspicion coefficient value range is [0, 1], and the size represents the possibility of the user having the abnormal electricity consumption behaviors; wherein, when the value of the suspicion coefficient is in the range of [0,0.5), the suspicion coefficient belongs to normal power utilization; when the value of the suspicion coefficient is in the range of [0.5,1], the power utilization is abnormal.
11. A power-consumption inspection intelligent diagnosis system using the power-consumption inspection intelligent diagnosis method according to any one of claims 1 to 10, comprising a data layer, an application layer and a presentation layer; the method is characterized in that:
the data layer comprises a data acquisition module and a data preprocessing module;
the application layer comprises an anti-electricity-stealing behavior identification model based on a fuzzy neural network, a user electricity consumption behavior comprehensive analysis module, a statistical analysis and report formulation module and an auxiliary information query module;
the display layer comprises a system management page, a data integration and query page and a power utilization abnormity analysis page.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110056298.8A CN112884000A (en) | 2021-01-15 | 2021-01-15 | Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110056298.8A CN112884000A (en) | 2021-01-15 | 2021-01-15 | Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112884000A true CN112884000A (en) | 2021-06-01 |
Family
ID=76048271
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110056298.8A Pending CN112884000A (en) | 2021-01-15 | 2021-01-15 | Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112884000A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117421610A (en) * | 2023-12-19 | 2024-01-19 | 山东德源电力科技股份有限公司 | Data anomaly analysis method for electric energy meter running state early warning |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866321A (en) * | 2012-08-13 | 2013-01-09 | 广东电网公司电力科学研究院 | Self-adaptive stealing-leakage prevention diagnosis method |
| CN108416693A (en) * | 2018-02-01 | 2018-08-17 | 东华大学 | A kind of multiplexing electric abnormality Activity recognition method based on fuzzy neural network |
| CN109741091A (en) * | 2018-12-18 | 2019-05-10 | 河海大学 | Customer charge classification method based on basic load abatement strategy |
| CN110097297A (en) * | 2019-05-21 | 2019-08-06 | 国网湖南省电力有限公司 | A kind of various dimensions stealing situation Intellisense method, system, equipment and medium |
| CN110223196A (en) * | 2019-06-04 | 2019-09-10 | 国网浙江省电力有限公司电力科学研究院 | Analysis method of opposing electricity-stealing based on typical industry feature database and sample database of opposing electricity-stealing |
| 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 |
| CN111738364A (en) * | 2020-08-05 | 2020-10-02 | 国网江西省电力有限公司供电服务管理中心 | Electricity stealing detection method based on combination of user load and electricity consumption parameter |
| CN112132210A (en) * | 2020-09-21 | 2020-12-25 | 国网辽宁省电力有限公司电力科学研究院 | Electricity stealing probability early warning analysis method based on customer electricity consumption behavior |
-
2021
- 2021-01-15 CN CN202110056298.8A patent/CN112884000A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102866321A (en) * | 2012-08-13 | 2013-01-09 | 广东电网公司电力科学研究院 | Self-adaptive stealing-leakage prevention diagnosis method |
| CN108416693A (en) * | 2018-02-01 | 2018-08-17 | 东华大学 | A kind of multiplexing electric abnormality Activity recognition method based on fuzzy neural network |
| CN109741091A (en) * | 2018-12-18 | 2019-05-10 | 河海大学 | Customer charge classification method based on basic load abatement strategy |
| CN110097297A (en) * | 2019-05-21 | 2019-08-06 | 国网湖南省电力有限公司 | A kind of various dimensions stealing situation Intellisense method, system, equipment and medium |
| CN110223196A (en) * | 2019-06-04 | 2019-09-10 | 国网浙江省电力有限公司电力科学研究院 | Analysis method of opposing electricity-stealing based on typical industry feature database and sample database of opposing electricity-stealing |
| 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 |
| CN111738364A (en) * | 2020-08-05 | 2020-10-02 | 国网江西省电力有限公司供电服务管理中心 | Electricity stealing detection method based on combination of user load and electricity consumption parameter |
| CN112132210A (en) * | 2020-09-21 | 2020-12-25 | 国网辽宁省电力有限公司电力科学研究院 | Electricity stealing probability early warning analysis method based on customer electricity consumption behavior |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117421610A (en) * | 2023-12-19 | 2024-01-19 | 山东德源电力科技股份有限公司 | Data anomaly analysis method for electric energy meter running state early warning |
| CN117421610B (en) * | 2023-12-19 | 2024-03-15 | 山东德源电力科技股份有限公司 | Data anomaly analysis method for electric energy meter running state early warning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110097297B (en) | Multi-dimensional electricity stealing situation intelligent sensing method, system, equipment and medium | |
| Li et al. | Electricity theft detection in power grids with deep learning and random forests | |
| CN110223196B (en) | Anti-electricity-stealing analysis method based on typical industry feature library and anti-electricity-stealing sample library | |
| CN107145966A (en) | Logic-based returns the analysis and early warning method of opposing electricity-stealing of probability analysis Optimized model | |
| CN106780115A (en) | Abnormal electricity consumption monitoring and alignment system and method | |
| CN109947815B (en) | Power theft identification method based on outlier algorithm | |
| CN112785458A (en) | Intelligent management and maintenance system for bridge health big data | |
| CN110636066B (en) | Network security threat situation assessment method based on unsupervised generative reasoning | |
| CN114676742A (en) | Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network | |
| CN111506635A (en) | System and method for analyzing residential electricity consumption behavior based on autoregressive naive Bayes algorithm | |
| CN110244099A (en) | Stealing detection method based on user's voltage | |
| CN115730962A (en) | Big data-based electric power marketing inspection analysis system and method | |
| CN113392910A (en) | Multi-dimensional intelligent analysis experience algorithm and classification algorithm for judging default electricity utilization and electricity stealing | |
| Zafar et al. | Step towards secure and reliable smart grids in Industry 5.0: A federated learning assisted hybrid deep learning model for electricity theft detection using smart meters | |
| CN112884000A (en) | Power utilization inspection intelligent diagnosis method and diagnosis system based on data mining | |
| CN111861587A (en) | System and method for analyzing residential electricity consumption behavior based on hidden Markov model and forward algorithm | |
| CN112418623A (en) | Anti-electricity-stealing identification method based on bidirectional long-time and short-time memory network and sliding window input | |
| CN116680641A (en) | User electricity consumption abnormality detection method based on machine learning algorithm | |
| CN111506636A (en) | System and method for analyzing residential electricity consumption behavior based on autoregressive and neighbor algorithm | |
| Lin et al. | Data-driven situation awareness of electricity-gas integrated energy system considering time series features | |
| Yao et al. | AdaBoost-CNN: a hybrid method for electricity theft detection | |
| CN114839462A (en) | Intelligent anti-electricity-stealing monitoring method and system | |
| Yang et al. | Anomaly detection of vehicle data based on LOF algorithm | |
| CN113435494A (en) | Low-voltage resident user abnormal electricity utilization identification method and simulation system | |
| CN113191409A (en) | Method for detecting abnormal electricity consumption behaviors of residents through tag data expansion and deep learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210601 |
|
| RJ01 | Rejection of invention patent application after publication |