CN109359848A - A kind of extremely relevant electricity consumer recognition methods of line loss and system - Google Patents
A kind of extremely relevant electricity consumer recognition methods of line loss and system Download PDFInfo
- Publication number
- CN109359848A CN109359848A CN201811171005.5A CN201811171005A CN109359848A CN 109359848 A CN109359848 A CN 109359848A CN 201811171005 A CN201811171005 A CN 201811171005A CN 109359848 A CN109359848 A CN 109359848A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- line loss
- electricity consumption
- feature vector
- 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.)
- Withdrawn
Links
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The present invention is suitable for power marketing field, provides a kind of extremely relevant electricity consumer recognition methods of line loss and system, wherein method includes: the electricity consumption data of user in the line loss data and platform area for obtain platform area；To data normalization processing, to obtain platform area line loss feature vector and user power consumption feature vector；Trend similarity factor and characteristic correlation coefficient are calculated according to described area's line loss feature vector and user power consumption feature vector；Multiplexing electric abnormality threshold value is calculated according to the electricity consumption data of history exception electricity consumption user and place platform area line loss data；Integrated correlation coefficient is calculated according to the trend similarity factor and characteristic correlation coefficient, judges whether the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, and then confirms whether user is multiplexing electric abnormality user.Integrated correlation coefficient provided in an embodiment of the present invention is determined jointly by trend similarity factor and characteristic correlation coefficient, it can effectively improve the accuracy rate of judgement, in addition, multiplexing electric abnormality threshold value is determined by history multiplexing electric abnormality user power consumption data, the reliability of multiplexing electric abnormality threshold value ensure that.
Description
Technical field
The present invention relates to power marketing field, more particularly to a kind of extremely relevant electricity consumer recognition methods of line loss be
System.
Background technique
With constantly improve for power equipment, electric power has also come into huge numbers of families.However, still having portion in current social
Point user carries out stealing using various means, this has not only influenced the line loss per unit technical indicator of Utilities Electric Co., while there is also
Serious security risk.
And in the prior art, for identification the method for user power consumption exception rely primarily on business personnel's many years from already
It tests and user's history electricity consumption data and platform area line loss per unit data artificial contrast is analyzed, this needs phase of the business personnel with many years
Pass experience, in addition, not can guarantee the accuracy rate for analyzing the result come yet, and needs are right from a large amount of user power consumption data
The abnormal electricity consumption than analyzing, needs to expend a large amount of manpower and energy.
As it can be seen that the extremely relevant electricity consumer recognition methods of existing line loss there is low efficiency, not can guarantee the standard of result
True rate needs to expend the technical issues of a large amount of manpower and energy, especially when needing to handle a large amount of user power consumption, makes
Manually such as also easily there are omitting at the technical problems in comparative analysis.
Summary of the invention
The embodiment of the present invention provides a kind of extremely relevant electricity consumer recognition methods of line loss and system, it is intended to solve existing skill
Low efficiency existing for art, needs to expend the technical issues of a large amount of manpower and energy at the accuracy rate that not can guarantee result.
The embodiment of the present invention provides a kind of line loss extremely relevant electricity consumer recognition methods, and the method includes following steps
It is rapid:
It obtains in multiple line loss data and corresponding time in platform area to be analyzed in preset time range in described area
Multiple electricity consumption data of all users；
Multiple electricity consumption data of multiple line loss data and user to described area are standardized, to obtain
Platform area line loss feature vector and user power consumption feature vector；
Trend similarity factor is calculated according to described area's line loss feature vector and user power consumption feature vector；
Characteristic correlation coefficient is calculated according to described area's line loss feature vector and user power consumption feature vector；
According to the electricity consumption data of history exception electricity consumption user and platform area where exception electricity consumption user described in the corresponding time
Line loss data calculate multiplexing electric abnormality threshold value；
Integrated correlation coefficient is calculated according to the trend similarity factor and characteristic correlation coefficient, and judges the comprehensive phase
Whether relationship number is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, whether confirmation user is multiplexing electric abnormality user.
The embodiment of the present invention also provides a kind of line loss extremely relevant electricity consumer identifying system, the system comprises:
Data acquisition unit, for obtaining multiple line loss data and correspondence in platform area to be analyzed in preset time range
In time in described area all users multiple electricity consumption data；
Data normalization processing unit, multiple electricity consumption numbers for multiple line loss data and user to described area
According to being standardized, to obtain platform area line loss feature vector and user power consumption feature vector；
Trend similarity factor computing unit, for according to described area's line loss feature vector and user power consumption feature vector
Calculating trend similarity factor；
Characteristic correlation coefficient computing unit, for according to described area's line loss feature vector and user power consumption feature vector
Calculate characteristic correlation coefficient；
Multiplexing electric abnormality threshold computation unit, for according to the electricity consumption data of history exception electricity consumption user in the corresponding time
Platform area's line loss data where the exception electricity consumption user calculate multiplexing electric abnormality threshold value；
Multiplexing electric abnormality judging unit is comprehensive related for being calculated according to the trend similarity factor and characteristic correlation coefficient
Coefficient, and judge whether the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, confirmation is used
Whether family is multiplexing electric abnormality user.
In multiple line loss data and corresponding time of the present invention by obtaining the area preset time Nei Tai in described area
Multiple electricity consumptions of all users, and the data are standardized, the data after standardization can be more
Adding intuitively indicates that the data the distance between to average data, then calculate trend by related algorithm to standardized data
Similarity factor and characteristic correlation coefficient, the trend similarity factor indicate user power consumption feature vector and platform area line loss feature
The similarity degree of each data variation trend in vector, the trend similarity factor is bigger, show more to meet user power consumption it is less,
The case where platform area line loss increases, illustrates that user is higher a possibility that there are multiplexing electric abnormalities, and the calculating of the characteristic correlation coefficient
Depend only on the relationship of ranking between data, and the size independent of each data, it can effectively improve the algorithm
Robustness, in addition, being determined that user may deposit by the data to history multiplexing electric abnormality data and corresponding platform area same period line loss
In the standard of abnormal electricity consumption, the accuracy rate of judgement is improved.The extremely relevant electricity consumer of line loss provided in an embodiment of the present invention is known
Other method and system substantially increase the efficiency of identification, and accuracy rate with higher and robustness.
Detailed description of the invention
Fig. 1 is a kind of overall flow figure of the extremely relevant electricity consumer recognition methods of line loss provided in an embodiment of the present invention；
Fig. 2 is the flow chart of standardization provided in an embodiment of the present invention；
Fig. 3 is the flow chart of calculating trend similarity factor provided in an embodiment of the present invention；
Fig. 4 is the flow chart provided in an embodiment of the present invention for calculating characteristic correlation coefficient；
Fig. 5 is the flow chart provided in an embodiment of the present invention for calculating multiplexing electric abnormality threshold value；
Fig. 6 is the structure chart of the extremely relevant electricity consumer identifying system of line loss provided in an embodiment of the present invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The embodiment of the invention provides a kind of extremely relevant electricity consumer recognition methods of line loss based on many algorithms, utilize
Many algorithms handle user power consumption data and platform area line loss data, can determine user power consumption data and Tai Qu
The integrated correlation coefficient of line loss data, and determine user power consumption with the presence or absence of abnormal, entire recognition methods standard with higher
True rate and robustness.The embodiment of the invention also provides a kind of extremely relevant electricity consumer identifications of line loss based on many algorithms
System is supported to obtain by corresponding equipment, calculates, handles user power consumption data and platform area line loss data, compared to existing
Have in technology and analysis comparison is carried out to electricity consumption data and platform area line loss data using manpower, work effect can be effectively improved
Rate, and the accuracy rate judged is higher, while also ing save a large amount of human cost.
Fig. 1 is a kind of extremely relevant electricity consumer recognition methods of line loss provided in an embodiment of the present invention, and details are as follows.
Step S101 obtains institute in multiple line loss data and corresponding time in platform area to be analyzed in preset time range
Multiple electricity consumption data of all users in the area Shu Tai.
In the embodiment of the present invention, multiple data in the preset time range can be the use of every month in nearest 1 year
Electricity data is also possible to the electricity consumption data in each week in nearest one month, can also be in a nearest week daily
Electricity consumption data.
As a preferred embodiment of the present invention, the preset range is the electricity consumption in each week in a nearest season
Measure data.
In embodiments of the present invention, the line loss data in described area and the electricity consumption data of user are an a pair in time
It answers, the line loss data in the area each time point Zhong Tai correspond to the electricity consumption data of a user.
Multiple electricity consumption data of step S102, multiple line loss data and user to described area are standardized place
Reason, to obtain platform area line loss feature vector and user power consumption feature vector.
In embodiments of the present invention, the data after standardization more intuitive can indicate the data to putting down
The distance between equal data, the interference of data amplitudes and extraneous factor can be effectively reduced by standardization, for example,
With the electricity consumption in some timing of a power consumer, and the line loss per unit in its corresponding platform area in the timing, in data amplitudes
Differ very big on (order of magnitude), but the data after standardization, ordinal number when still ensuring that the electricity consumption of single user
It is comparable on amplitude according to the time series data with platform area line loss.
In embodiments of the present invention, due to multiple line loss data in described area and multiple electricity consumption data of the user
It is one-to-one relationship in time, therefore special by the area standardization Hou Tai line loss feature vector and user power consumption
Each data still maintain one-to-one relationship in time in sign vector.
Step S103, according to described area's line loss feature vector system similar to user power consumption feature vector calculating trend
Number.
In embodiments of the present invention, the trend similarity factor indicates user power consumption feature vector and platform area line loss feature
The similarity degree of each data variation trend in vector, the trend similarity factor is bigger, show more to meet " user power consumption compared with
Few, platform area line loss increases " the case where, illustrate that user is higher a possibility that there are multiplexing electric abnormalities.
Step S104 calculates feature phase relation according to described area's line loss feature vector and user power consumption feature vector
Number.
In embodiments of the present invention, the calculating of the characteristic correlation coefficient depends only on the relationship of ranking between data,
And therefore the size independent of each data can effectively improve the robustness of the algorithm, so that the result of judgement is more
Accurately.
Step S105, according to the electricity consumption data of history exception electricity consumption user and exception electricity consumption user described in the corresponding time
Place platform area line loss data calculate multiplexing electric abnormality threshold value.
In embodiments of the present invention, it is determined by the data to history multiplexing electric abnormality data and corresponding platform area same period line loss
There may be the standard of abnormal electricity consumption by user, improves the accuracy rate of judgement.
In embodiments of the present invention, it needs to the electricity consumption data of the history exception electricity consumption user and institute in the corresponding time
Platform area line loss data where stating abnormal electricity consumption user, which are standardized, generates abnormal user electricity consumption feature vector and different
Platform area line loss feature vector where common family, and successively calculate abnormal user electricity consumption feature vector and abnormal user place platform area
Trend similarity factor and characteristic correlation coefficient between line loss feature vector.
In embodiments of the present invention, described that multiplexing electric abnormality is calculated according to the trend similarity factor and characteristic correlation coefficient
The method of threshold value includes by the trend similarity factor and characteristic correlation coefficient weighting summation, calculates the trend similarity factor
And the breadth coefficient of characteristic correlation coefficient, it can also be modified according to related service experience in the case where combining actual conditions.
Step S106 obtains integrated correlation coefficient according to the trend similarity factor and characteristic correlation coefficient, and judges
Whether the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, whether confirmation user is use
Electrical anomaly user.
In embodiments of the present invention, when judging that the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, confirm institute
Stating user is multiplexing electric abnormality user；When judging that the integrated correlation coefficient is less than the multiplexing electric abnormality threshold value, described in confirmation
User is electricity consumption normal users.
In multiple line loss data and corresponding time of the present invention by obtaining the area preset time Nei Tai in described area
Multiple electricity consumptions of all users, and the data are standardized, the data after standardization can be more
Adding intuitively indicates that the data the distance between to average data, then calculate trend by related algorithm to standardized data
Similarity factor and characteristic correlation coefficient, the trend similarity factor indicate user power consumption feature vector and platform area line loss feature
The similarity degree of each data variation trend in vector, the trend similarity factor is bigger, show more to meet user power consumption it is less,
The case where platform area line loss increases, illustrates that user is higher a possibility that there are multiplexing electric abnormalities, and the calculating of the characteristic correlation coefficient
Depend only on the relationship of ranking between data, and the size independent of each data, it can effectively improve the algorithm
Robustness, in addition, being determined that user may deposit by the data to history multiplexing electric abnormality data and corresponding platform area same period line loss
In the standard of abnormal electricity consumption, the accuracy rate of judgement is improved.The extremely relevant electricity consumer of line loss provided in an embodiment of the present invention is known
Other method and system substantially increase the efficiency of identification, and accuracy rate with higher and robustness.
Fig. 2 is standardized flow chart provided in an embodiment of the present invention, and details are as follows.
Step S201 generates platform area line loss to described area's line loss data-conversion and negates data.
In embodiments of the present invention, since platform area line loss data and user power consumption data are negatively correlated, by described area
Line loss data-conversion enables to platform area line loss to negate data to be positively correlated with user power consumption data, is convenient for subsequent processing.
Step S202 calculates described area's line loss and negates being averaged for the average of data and the electricity consumption data of user
Number.
Step S203 calculates the standard that described area's line loss negates the standard deviation of data and the electricity consumption data of user
Difference.
Step S204, the multiple line losses for calculating described area negate data and described area's line loss negates the average of data
Line loss negate data difference, and calculate the line loss and negate the line that data difference negates the standard deviation of data divided by described area's line loss
Damage negates data quotient.
Step S205 calculates the average of multiple electricity consumption data of the user and the electricity consumption data of the user
Electricity consumption data difference, and the electricity consumption data difference is calculated divided by the electricity consumption number of the standard deviation of the electricity consumption data of the user
According to quotient.
In embodiments of the present invention, above-mentioned steps S202~step S205 is to be standardized place using criterion score algorithm
The detailed process of reason, the specific formula such as following formula of the criterion score algorithm:
Z=(x- μ)/σ
Wherein, the μ indicates that the average of vector, σ indicate the standard deviation of data.
Step S206 negates data quotient according to the multiple line loss and generates platform area line loss feature vector.
Step S207 generates user power consumption feature vector according to the multiple electricity consumption data quotient.
In embodiments of the present invention, due to multiple line loss data in described area and multiple electricity consumption data of the user
It is one-to-one relationship in time, therefore special by the area standardization Hou Tai line loss feature vector and user power consumption
Each data still maintain one-to-one relationship in time in sign vector.
In embodiments of the present invention, the data after standardization more intuitive can indicate the data to putting down
The distance between equal data, the interference of data amplitudes and extraneous factor can be effectively reduced by standardization, for example,
With the electricity consumption in some timing of a power consumer, and the line loss per unit in its corresponding platform area in the timing, in data amplitudes
Differ very light on (order of magnitude), but the data after standardization, ordinal number when still ensuring that the electricity consumption of single user
It is comparable on amplitude according to the time series data with platform area line loss.
Fig. 3 is the flow chart of calculating trend similarity factor provided in an embodiment of the present invention, and details are as follows.
Each data in described area's line loss feature vector are compared by step S301 with a upper data, are generated
Electricity consumption identification strings.
As an embodiment of the present invention, by each data and a upper data in described area's line loss feature vector
It is compared, "+" is labeled as if rising, "-" is labeled as if decline, generate the line loss being only made of "+" and/or "-"
Rate identification strings.
Each data in the user power consumption feature vector are compared by step S302 with a upper data, are generated
Line loss identification strings.
As an embodiment of the present invention, by each data in the user power consumption feature vector and a upper data
It is compared, "+" is labeled as if rising, "-" is labeled as if decline, generate the line loss being only made of "+" and/or "-"
Rate identification strings.
In embodiments of the present invention, due in platform area line loss feature vector and user power consumption feature vector each data when
Between on still maintain one-to-one relationship, therefore the electricity consumption identification strings and line loss per unit identification strings are in the time
On equally remain one-to-one relationship.
It is identical to calculate corresponding position character in the electricity consumption identification strings and line loss identification strings by step S303
Character account for the proportionality coefficients of entire identification strings, the proportionality coefficient is trend similarity factor.
In the embodiment of the present invention, the trend similarity factor indicate user power consumption feature vector and platform area line loss feature to
The similarity degree of each data variation trend in amount, the trend similarity factor is bigger, shows more to meet user power consumption less, platform
The case where area's line loss increases, illustrates that user is higher a possibility that there are multiplexing electric abnormalities.
Fig. 4 is the flow chart provided in an embodiment of the present invention for calculating characteristic correlation coefficient, and details are as follows.
Step S401 obtains the sequence serial number of each data and user in described area's line loss feature vector according to size relation
The sequence serial number of each data in electricity consumption feature vector.
In embodiments of the present invention, the size relation includes sorting and sorting from large to small from small to large, but need to protect
It is identical as the sortord of user power consumption feature vector to hold described area's line loss feature vector.
Step S402 is calculated in platform area line loss feature vector in the sequence serial number and user power consumption feature vector of each data
Square of the difference of the sequence serial number of corresponding data.
Step S403, square summation to the difference of the sequence serial number of each data, and calculate characteristic correlation coefficient.
In embodiments of the present invention, the characteristic correlation coefficient η is calculated by following formula:
η=1-6A/n3-n
Wherein, A is the quadratic sum of the difference of the sequence serial number of each data, and n is number in described area's line loss feature vector
According to number.
In embodiments of the present invention, above-mentioned steps S202~step S205 is to be calculated using Spearman rank correlation algorithm
The detailed process of characteristic correlation coefficient, the process of Spearman rank correlation algorithm for ease of description, by taking following examples as an example:
If user power consumption feature vector be (0, -0.5,1,0.5, -1), platform area line loss feature vector be (- 10%, -
12%, -9.5%, -11%, -11.5%), then according to sorting from small to large, the sequence sequence of user power consumption feature vector can be obtained
It number is (3,2,5,4,1), the sequence serial number (4,1,5,3,2) of platform area line loss feature vector takes the arrangement serial number of corresponding data
Difference quadratic sum be (3-4)2+(2-1)2+(5-5)2+(4-3)2+(1-2)2=4, then relative coefficient η is calculated as according to formula
0.8。
In embodiments of the present invention, due to the relationship between the not instead of data of Spearman rank algorithm checks, number
According to the relationship between ranking, there is stronger robustness using the calculated result of Spearman rank algorithm.
Fig. 5 is the flow chart provided in an embodiment of the present invention for calculating multiplexing electric abnormality threshold value, and details are as follows.
Step S501 uses the electricity consumption data of the history exception electricity consumption user with exception electricity consumption described in the corresponding time
Platform area line loss data are standardized where family.
In embodiments of the present invention, described different in the corresponding time to the electricity consumption data of the history exception electricity consumption user
Platform area line loss data where common electricity user are standardized, with obtain the electricity consumption measure feature of history exception electricity consumption user to
Amount and platform area line loss feature vector where abnormal electricity consumption user.
In embodiments of the present invention, the standardization is to be standardized using criterion score algorithm.
Step S502, according to the electricity consumption feature vector of the history exception electricity consumption user and the abnormal electricity consumption user institute
In platform area, line loss feature vector calculates the trend similarity factor of abnormal electricity consumption user.
Step S503, according to the electricity consumption feature vector of the history exception electricity consumption user and the abnormal electricity consumption user institute
In platform area, line loss feature vector calculates the characteristic correlation coefficient of abnormal electricity consumption user.
In embodiments of the present invention, history exception electricity consumption user is judged using trend similarity factor and characteristic correlation coefficient
Electricity consumption feature vector and the abnormal electricity consumption user where there may be relationships between platform area line loss feature vector.
In embodiments of the present invention, by combining trend similarity factor and characteristic correlation coefficient, the standard of judgement is improved
True rate.
Step S504, according to the feature phase of the trend similarity factor of the abnormal electricity consumption user and the abnormal electricity consumption user
Relationship number determines multiplexing electric abnormality threshold value.
In embodiments of the present invention, described that multiplexing electric abnormality is calculated according to the trend similarity factor and characteristic correlation coefficient
The method of threshold value includes by the trend similarity factor and characteristic correlation coefficient weighting summation, calculates the trend similarity factor
And the breadth coefficient of characteristic correlation coefficient, it can also be modified according to related service experience in the case where combining actual conditions.
Fig. 6 is the structure chart of the extremely relevant electricity consumer identifying system of line loss provided in an embodiment of the present invention, for the ease of
Illustrate, part related to the embodiment of the present invention is only shown.
In embodiments of the present invention, the extremely relevant electricity consumer identifying system of the line loss includes:
Data acquisition unit 601, for obtain platform area to be analyzed in preset time range multiple line loss data and
Correspond to multiple electricity consumption data of all users in described area in the time.
In the embodiment of the present invention, multiple electricity consumption data in the preset time range can be each in nearest 1 year
The electricity consumption data of the moon, are also possible to the electricity consumption data in each week in nearest one month, can also be a nearest week
Interior daily electricity consumption data.
As a preferred embodiment of the present invention, the preset range is the electricity consumption in each week in a nearest season
Measure data.
In embodiments of the present invention, the line loss data in described area and the electricity consumption data of user are an a pair in time
It answers, the line loss data in the area each time point Zhong Tai correspond to the electricity consumption data of a user.
Data normalization processing unit 602, multiple electricity consumptions for multiple line loss data and user to described area
Amount data are standardized, to obtain platform area line loss feature vector and user power consumption feature vector.
In embodiments of the present invention, the data after standardization more intuitive can indicate the data to putting down
The distance between equal data, the interference of data amplitudes and extraneous factor can be effectively reduced by standardization, for example,
With the electricity consumption in some timing of a power consumer, and the line loss per unit in its corresponding platform area in the timing, in data amplitudes
Differ very big on (order of magnitude), but the data after standardization, ordinal number when still ensuring that the electricity consumption of single user
It is comparable on amplitude according to the time series data with platform area line loss.
In embodiments of the present invention, due to multiple line loss data in described area and multiple electricity consumption data of the user
It is one-to-one relationship in time, therefore special by the area standardization Hou Tai line loss feature vector and user power consumption
Each data still maintain one-to-one relationship in time in sign vector.
Trend similarity factor computing unit 603, for according to described area's line loss feature vector and user power utilization measure feature
Vector calculates trend similarity factor.
In embodiments of the present invention, the trend similarity factor indicates user power consumption feature vector and platform area line loss feature
The similarity degree of each data variation trend in vector, the trend similarity factor is bigger, show more to meet user power consumption it is less,
The case where platform area line loss increases, illustrates that user is higher a possibility that there are multiplexing electric abnormalities.
Characteristic correlation coefficient computing unit 604, for according to described area's line loss feature vector and user power utilization measure feature
Vector calculates characteristic correlation coefficient.
In embodiments of the present invention, the calculating of the characteristic correlation coefficient depends only on the relationship of ranking between data,
And therefore the size independent of each data can effectively improve the robustness of the algorithm, so that the result of judgement is more
Accurately.
Multiplexing electric abnormality threshold computation unit 605, for according to the electricity consumption data of history exception electricity consumption user with to it is corresponding when
Platform area's line loss data where the interior abnormal electricity consumption user calculate multiplexing electric abnormality threshold value.
In embodiments of the present invention, it is determined by the data to history multiplexing electric abnormality data and corresponding platform area same period line loss
There may be the standard of abnormal electricity consumption by user, improves the accuracy rate of judgement.
In embodiments of the present invention, it needs to the electricity consumption data of the history exception electricity consumption user and institute in the corresponding time
Platform area line loss data where stating abnormal electricity consumption user, which are standardized, generates abnormal user electricity consumption feature vector and different
Platform area line loss feature vector where common family, and successively calculate abnormal user electricity consumption feature vector and abnormal user place platform area
Trend similarity factor and characteristic correlation coefficient between line loss feature vector.
In embodiments of the present invention, described that multiplexing electric abnormality is calculated according to the trend similarity factor and characteristic correlation coefficient
The method of threshold value includes by the trend similarity factor and characteristic correlation coefficient weighting summation, calculates the trend similarity factor
And the breadth coefficient of characteristic correlation coefficient, it can also be modified according to related service experience in the case where combining actual conditions.
Multiplexing electric abnormality judging unit 606, for calculating synthesis according to the trend similarity factor and characteristic correlation coefficient
Related coefficient, and judge whether the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, true
Recognize whether user is multiplexing electric abnormality user.
In embodiments of the present invention, when judging that the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, confirm institute
Stating user is multiplexing electric abnormality user；When judging that the integrated correlation coefficient is less than the multiplexing electric abnormality threshold value, described in confirmation
User is electricity consumption normal users.
In multiple line loss data and corresponding time of the present invention by obtaining the area preset time Nei Tai in described area
Multiple electricity consumptions of all users, and the data are standardized, the data after standardization can be more
Adding intuitively indicates that the data the distance between to average data, then calculate trend by related algorithm to standardized data
Similarity factor and characteristic correlation coefficient, the trend similarity factor indicate user power consumption feature vector and platform area line loss feature
The similarity degree of each data variation trend in vector, the trend similarity factor is bigger, show more to meet user power consumption it is less,
The case where platform area line loss increases, illustrates that user is higher a possibility that there are multiplexing electric abnormalities, and the calculating of the characteristic correlation coefficient
Depend only on the relationship of ranking between data, and the size independent of each data, it can effectively improve the algorithm
Robustness, in addition, being determined that user may deposit by the data to history multiplexing electric abnormality data and corresponding platform area same period line loss
In the standard of abnormal electricity consumption, the accuracy rate of judgement is improved.The extremely relevant electricity consumer of line loss provided in an embodiment of the present invention is known
Other method and system substantially increase the efficiency of identification, and accuracy rate with higher and robustness.
Claims (7)
1. a kind of extremely relevant electricity consumer recognition methods of line loss, which is characterized in that the described method comprises the following steps:
It obtains in multiple line loss data and corresponding time in platform area to be analyzed in preset time range and owns in described area
Multiple electricity consumption data of user；
Multiple electricity consumption data of multiple line loss data and user to described area are standardized, to obtain platform area
Line loss feature vector and user power consumption feature vector；
Trend similarity factor is calculated according to described area's line loss feature vector and user power consumption feature vector；
Characteristic correlation coefficient is calculated according to described area's line loss feature vector and user power consumption feature vector；
According to the electricity consumption data of history exception electricity consumption user and platform area line loss where exception electricity consumption user described in the corresponding time
Data calculate multiplexing electric abnormality threshold value；
Integrated correlation coefficient is calculated according to the trend similarity factor and characteristic correlation coefficient, and judges the comprehensive phase relation
Whether number is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, whether confirmation user is multiplexing electric abnormality user.
2. the method according to claim 1, wherein the multiple line loss data and user to described area
Multiple electricity consumption data be standardized, to obtain the step of platform area line loss feature vector and user power consumption feature vector
Suddenly, it specifically includes:
Platform area line loss is generated to described area's line loss data-conversion and negates data；
Calculate the average that described area's line loss negates the average of data and the electricity consumption data of user；
Calculate the standard deviation that described area's line loss negates the standard deviation of data and the electricity consumption data of user；
The line loss for the average that the multiple line losses for calculating described area negate data and described area's line loss negates data negates number
According to difference, and calculates the line loss and negate the line loss of the standard deviation that data difference negates data divided by described area's line loss and negate data
Quotient；
The electricity consumption data difference of the average of multiple electricity consumption data of the user and the electricity consumption data of the user is calculated,
And the electricity consumption data difference is calculated divided by the electricity consumption data quotient of the standard deviation of the electricity consumption data of the user；
Data quotient, which is negated, according to the multiple line loss generates platform area line loss feature vector；
User power consumption feature vector, the user power consumption feature vector and institute are generated according to the multiple electricity consumption data quotient
Each data are one-to-one relationship in the area Shu Tai line loss feature vector.
3. the method according to claim 1, wherein described use according to described area's line loss feature vector and user
Electricity feature vector calculates the step of trend similarity factor, specifically includes:
Each data in described area's line loss feature vector are compared with a upper data, electricity consumption is generated and identifies character
String；
Each data in the user power consumption feature vector are compared with a upper data, line loss is generated and identifies character
String；
The identical character of corresponding position character in the electricity consumption identification strings and line loss identification strings is calculated to account for entirely
The proportionality coefficient of identification strings, the proportionality coefficient are trend similarity factor.
4. the method according to claim 1, wherein according to described area's line loss feature vector and user power consumption
Feature vector calculates the step of characteristic correlation coefficient, specifically includes:
According to size relation obtain in described area's line loss feature vector the sequence serial number of each data and user power utilization measure feature to
The sequence serial number of each data in amount；
Calculate the row of corresponding data in the sequence serial number and user power consumption feature vector of each data in platform area line loss feature vector
Square of the difference of sequence serial number；
Square summation to the difference of the sequence serial number of each data, and characteristic correlation coefficient is calculated, the characteristic correlation coefficient η
It is calculated by following formula:
η=1-6An/3-n
Wherein, A is the quadratic sum of the difference of the sequence serial number of each data, and n is data in described area's line loss feature vector
Number.
5. the method according to claim 1, wherein according to the electricity consumption data of history exception electricity consumption user with it is right
The step of platform area's line loss data where the seasonable interior abnormal electricity consumption user calculate multiplexing electric abnormality threshold value, specifically includes:
To the electricity consumption data of the history exception electricity consumption user and platform area line where exception electricity consumption user described in the corresponding time
Damage data are standardized, to obtain electricity consumption feature vector and the abnormal electricity consumption user place of history exception electricity consumption user
Platform area line loss feature vector
It is special according to platform area line loss where the electricity consumption feature vector of the history exception electricity consumption user and the abnormal electricity consumption user
Levy the trend similarity factor that vector calculates abnormal electricity consumption user；
It is special according to platform area line loss where the electricity consumption feature vector of the history exception electricity consumption user and the abnormal electricity consumption user
Levy the characteristic correlation coefficient that vector calculates abnormal electricity consumption user；
It is determined and is used according to the characteristic correlation coefficient of the trend similarity factor of the abnormal electricity consumption user and the abnormal electricity consumption user
Electrical anomaly threshold value.
6. the method according to claim 1, wherein it is described according to the judgement as a result, confirmation user whether
It the step of for multiplexing electric abnormality user, specifically includes:
When judging that the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, confirm that the user is multiplexing electric abnormality user；
When judging that the integrated correlation coefficient is less than the multiplexing electric abnormality threshold value, confirm that the user is that electricity consumption is just common
Family.
7. a kind of extremely relevant electricity consumer identifying system of line loss, which is characterized in that the system comprises:
Data acquisition unit, for obtaining multiple line loss data in platform area to be analyzed in preset time range and corresponding to the time
Multiple electricity consumption data of all users in interior described area；
Data normalization processing unit, for multiple line loss data and user to described area multiple electricity consumption data into
Row standardization, to obtain platform area line loss feature vector and user power consumption feature vector；
Trend similarity factor computing unit, for being calculated according to described area's line loss feature vector and user power consumption feature vector
Trend similarity factor；
Characteristic correlation coefficient computing unit, for being calculated according to described area's line loss feature vector and user power consumption feature vector
Characteristic correlation coefficient；
Multiplexing electric abnormality threshold computation unit, for described in the corresponding time according to the electricity consumption data of history exception electricity consumption user
Platform area's line loss data calculate multiplexing electric abnormality threshold value where abnormal electricity consumption user；
Multiplexing electric abnormality judging unit, for calculating comprehensive phase relation according to the trend similarity factor and characteristic correlation coefficient
Number, and judge whether the integrated correlation coefficient is more than the multiplexing electric abnormality threshold value, according to the judgement as a result, confirmation user
It whether is multiplexing electric abnormality user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811171005.5A CN109359848A (en) | 2018-10-09 | 2018-10-09 | A kind of extremely relevant electricity consumer recognition methods of line loss and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811171005.5A CN109359848A (en) | 2018-10-09 | 2018-10-09 | A kind of extremely relevant electricity consumer recognition methods of line loss and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109359848A true CN109359848A (en) | 2019-02-19 |
Family
ID=65348723
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811171005.5A Withdrawn CN109359848A (en) | 2018-10-09 | 2018-10-09 | A kind of extremely relevant electricity consumer recognition methods of line loss and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109359848A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110472871A (en) * | 2019-08-16 | 2019-11-19 | 广东电网有限责任公司 | A kind of investigation method for management line loss abnormal cause |
| CN110503136A (en) * | 2019-07-31 | 2019-11-26 | 国家电网有限公司 | Platform area line loss exception analysis method, computer readable storage medium and terminal device |
| CN110988422A (en) * | 2019-12-19 | 2020-04-10 | 北京中电普华信息技术有限公司 | Electricity stealing identification method and device and electronic equipment |
| CN111062620A (en) * | 2019-12-19 | 2020-04-24 | 烟台海颐软件股份有限公司 | Intelligent analysis system and method for electric power charging fairness based on hybrid charging data |
| CN111521868A (en) * | 2020-04-28 | 2020-08-11 | 广东电网有限责任公司梅州供电局 | Method and device for screening electricity stealing users based on big metering data |
| CN112418687A (en) * | 2020-11-26 | 2021-02-26 | 广东电网有限责任公司 | User electricity utilization abnormity identification method and device based on electricity utilization characteristics and storage medium |
| CN112816774A (en) * | 2020-12-15 | 2021-05-18 | 国网宁夏电力有限公司营销服务中心(国网宁夏电力有限公司计量中心) | Electricity stealing troubleshooting method based on big data |
| CN112861128A (en) * | 2021-01-21 | 2021-05-28 | 微梦创科网络科技（中国）有限公司 | Method and system for identifying machine accounts in batches |
| CN113112177A (en) * | 2021-04-28 | 2021-07-13 | 国网山东省电力公司泰安供电公司 | Transformer area line loss processing method and system based on mixed indexes |
| CN113253014A (en) * | 2021-04-07 | 2021-08-13 | 国网河北省电力有限公司衡水供电分公司 | Method, device and equipment for detecting abnormal topological relation of transformer area subscriber |
-
2018
- 2018-10-09 CN CN201811171005.5A patent/CN109359848A/en not_active Withdrawn
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110503136A (en) * | 2019-07-31 | 2019-11-26 | 国家电网有限公司 | Platform area line loss exception analysis method, computer readable storage medium and terminal device |
| CN110472871A (en) * | 2019-08-16 | 2019-11-19 | 广东电网有限责任公司 | A kind of investigation method for management line loss abnormal cause |
| CN110472871B (en) * | 2019-08-16 | 2023-07-07 | 广东电网有限责任公司 | Investigation method for managing line loss abnormal reasons |
| CN110988422B (en) * | 2019-12-19 | 2022-04-26 | 北京中电普华信息技术有限公司 | Electricity stealing identification method and device and electronic equipment |
| CN110988422A (en) * | 2019-12-19 | 2020-04-10 | 北京中电普华信息技术有限公司 | Electricity stealing identification method and device and electronic equipment |
| CN111062620A (en) * | 2019-12-19 | 2020-04-24 | 烟台海颐软件股份有限公司 | Intelligent analysis system and method for electric power charging fairness based on hybrid charging data |
| CN111062620B (en) * | 2019-12-19 | 2022-07-29 | 烟台海颐软件股份有限公司 | Intelligent electric power charging fairness analysis system and method based on hybrid charging data |
| CN111521868A (en) * | 2020-04-28 | 2020-08-11 | 广东电网有限责任公司梅州供电局 | Method and device for screening electricity stealing users based on big metering data |
| CN112418687A (en) * | 2020-11-26 | 2021-02-26 | 广东电网有限责任公司 | User electricity utilization abnormity identification method and device based on electricity utilization characteristics and storage medium |
| CN112816774A (en) * | 2020-12-15 | 2021-05-18 | 国网宁夏电力有限公司营销服务中心(国网宁夏电力有限公司计量中心) | Electricity stealing troubleshooting method based on big data |
| CN112816774B (en) * | 2020-12-15 | 2023-01-06 | 国网宁夏电力有限公司营销服务中心(国网宁夏电力有限公司计量中心) | Electricity stealing troubleshooting method based on big data |
| CN112861128A (en) * | 2021-01-21 | 2021-05-28 | 微梦创科网络科技（中国）有限公司 | Method and system for identifying machine accounts in batches |
| CN113253014A (en) * | 2021-04-07 | 2021-08-13 | 国网河北省电力有限公司衡水供电分公司 | Method, device and equipment for detecting abnormal topological relation of transformer area subscriber |
| CN113112177A (en) * | 2021-04-28 | 2021-07-13 | 国网山东省电力公司泰安供电公司 | Transformer area line loss processing method and system based on mixed indexes |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109359848A (en) | A kind of extremely relevant electricity consumer recognition methods of line loss and system | |
| Westermann et al. | Unsupervised learning of energy signatures to identify the heating system and building type using smart meter data | |
| Filzmoser | A multivariate outlier detection method | |
| US8892376B2 (en) | Data processing device, data processing method, and program | |
| Beckel et al. | Towards automatic classification of private households using electricity consumption data | |
| CN107220906A (en) | Multiple Time Scales multiplexing electric abnormality analysis method based on electricity consumption acquisition system | |
| EP2595098A2 (en) | Method and system for detecting an appliance based on users' feedback information | |
| TWI829735B (en) | Computer system, computer readable storage medium and method for analysis of data from smart meters | |
| CN110416995B (en) | Non-invasive load decomposition method and device | |
| CN110147871A (en) | A kind of stealing detection method and system based on SOM neural network Yu K- mean cluster | |
| CN112418687B (en) | User electricity utilization abnormity identification method and device based on electricity utilization characteristics and storage medium | |
| CN110047013B (en) | Anti-private-transformer user intermittent electricity stealing method | |
| CN107247962B (en) | A kind of real-time electrical appliance recognition and system based on sliding window | |
| JP2006349519A (en) | Demand factor analyzing system and demand factor analysis method | |
| TWI391675B (en) | Method and system for estimating use time of electric appliances, and computer program product thereof | |
| Grigoraş et al. | Load estimation for distribution systems using clustering techniques | |
| Welikala et al. | A real-time non-intrusive load monitoring system | |
| CN114266457A (en) | Method for detecting different loss inducement of distribution line | |
| CN106779133B (en) | Household electricity consumption data prediction method | |
| CN111539840A (en) | Electricity stealing detection method and system fusing clustering and density estimation | |
| CN111126449A (en) | Battery fault classification diagnosis method based on cluster analysis | |
| CN110969539B (en) | Photovoltaic electricity stealing discovery method and system based on curve morphology analysis | |
| CN113642645A (en) | Low-voltage user phase identification method based on fuzzy C mean value and Hausdorff distance | |
| CN112132442A (en) | Method for evaluating load identification effect under intermittent start-stop electric heating working condition | |
| Nuran et al. | Non-Intrusive Load Monitoring Method for Appliance Identification Using Random Forest Algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190219 |
|
| WW01 | Invention patent application withdrawn after publication |