CN108460410A - Electricity consumption mode identification method and system, the storage medium of citizen requirement side - Google Patents
Electricity consumption mode identification method and system, the storage medium of citizen requirement side Download PDFInfo
- Publication number
- CN108460410A CN108460410A CN201810130203.0A CN201810130203A CN108460410A CN 108460410 A CN108460410 A CN 108460410A CN 201810130203 A CN201810130203 A CN 201810130203A CN 108460410 A CN108460410 A CN 108460410A
- Authority
- CN
- China
- Prior art keywords
- data
- power load
- data set
- load data
- user
- 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
- 230000005611 electricity Effects 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 34
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000006116 polymerization reaction Methods 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 8
- 230000009467 reduction Effects 0.000 claims description 12
- 240000002853 Nelumbo nucifera Species 0.000 claims description 9
- 235000006508 Nelumbo nucifera Nutrition 0.000 claims description 9
- 235000006510 Nelumbo pentapetala Nutrition 0.000 claims description 9
- 238000003909 pattern recognition Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 4
- 238000003064 k means clustering Methods 0.000 claims description 3
- 210000003127 knee Anatomy 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 235000021152 breakfast Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000002354 daily effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000001755 vocal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Public Health (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (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
The present invention provides a kind of the electricity consumption mode identification method and system, storage medium of citizen requirement side, is related to multiplexe electric technology field.This method includes the power load data obtained in default historical time section；The first data set is generated according to power load data；Dimension-reduction treatment is carried out respectively using each row power load data in the first data set of stage feeding polymerization approximate data pair, obtains the second data set；According to the second data set, initial cluster center is determined；It is clustered using each row power load data in the second data set of default clustering algorithm pair, obtains user power utilization pattern；Or generate third data set according to power load data；Each row power load data in third data set are clustered using the clustering algorithm based on dynamic time warping, obtain user power utilization pattern；Feature extraction is carried out to user power utilization pattern, obtains corresponding electricity consumption pattern feature.The present invention can be with the large-scale power load data of efficient process, and can improve cluster efficiency.
Description
Technical field
The present invention relates to multiplexe electric technology fields, and in particular to a kind of electricity consumption mode identification method of citizen requirement side and is
System, storage medium.
Background technology
As the improvement of people's living standards, residential electricity consumption is constantly increasing, the importance of residential electricity consumption demand side management
It is more prominent.Implement demand side management to be conducive to improve power supply reliability, integrates regenerative resource and access power grid, contribute to simultaneously
Resident reduces electric cost and improves efficiency of energy utilization.The identification of household electricity pattern can allow resident to be best understood from them
Electricity consumption situation, to reasonably optimizing electricity consumption with improve efficiency of energy utilization and reduce electric cost, this to Utilities Electric Co. implement
It is more efficiently also of great significance with flexible demand response management, it is therefore necessary to which a kind of side of electricity consumption pattern-recognition is provided
Case.
Invention content
(1) the technical issues of solving
In view of the deficiencies of the prior art, the present invention provides a kind of electricity consumption mode identification method of citizen requirement side and it is
System, storage medium can improve recognition efficiency with the large-scale power load data of efficient process.
(2) technical solution
In order to achieve the above object, the present invention is achieved by the following technical programs：
In a first aspect, the embodiment of the present invention provides a kind of electricity consumption mode identification method of citizen requirement side, including：
Obtain the power load data in default historical time section；
The first data set is generated according to the power load data；Wherein, the line number Yu the use of first data set
The corresponding user's number of electric load data is identical, and the i-th row includes each use of i-th of user in the default historical time section
Electric load data；Each row power load data in first data set are dropped respectively using stage feeding polymerization approximate data
Dimension processing, obtains the second data set；According to second data set, initial cluster center is determined；Based in the initial clustering
The heart clusters each row power load data in second data set using default clustering algorithm, obtains user power utilization
Pattern；
Alternatively, generating third data set according to the power load data；Wherein, the line number of the third data set is institute
The product of the number for the time interval that the corresponding user's number of power load data includes with the default historical time section is stated,
Power load data of each user in each time interval generate a line power load in the third data set
Data, the default historical time section includes multiple time intervals；Using the clustering algorithm based on dynamic time warping to institute
Each row power load data stated in third data set are clustered, and user power utilization pattern is obtained；
Feature extraction is carried out to the user power utilization pattern, obtains corresponding electricity consumption pattern feature.
Second aspect, the embodiment of the present invention provide a kind of citizen requirement side electricity consumption pattern recognition system, including data acquisition
Module, the first identification module or the second identification module and characteristic extracting module, wherein：
The data acquisition module is used to obtain the power load data in default historical time section；First identification
Module includes the first generation unit, Data Dimensionality Reduction unit, center determination unit and the first cluster cell；First generation unit
For generating the first data set according to the power load data；Wherein, the line number of first data set and the electricity consumption are negative
The corresponding user's number of lotus data is identical, and the i-th row includes that each electricity consumption of i-th of user in the default historical time section is negative
Lotus data；The Data Dimensionality Reduction unit is used for negative to each row electricity consumption in first data set using stage feeding polymerization approximate data
Lotus data carry out dimension-reduction treatment respectively, obtain the second data set；The center determination unit is used for according to second data set,
Determine initial cluster center；First cluster cell is used to be based on the initial cluster center, using default clustering algorithm pair
Each row power load data in second data set are clustered, and user power utilization pattern is obtained；
Second identification module includes the second generation unit and the second cluster cell；Second generation unit is used for according to institute
It states power load data and generates third data set；Wherein, the line number of the third data set corresponds to for the power load data
User's number and the default historical time section time interval that includes number product, each user is at each
Power load data in time interval generate a line power load data in the third data set, when the default history
Between section include multiple time intervals；Second cluster cell is used for using the clustering algorithm based on dynamic time warping to institute
Each row power load data stated in third data set are clustered, and user power utilization pattern is obtained；
The characteristic extracting module is used to carry out feature extraction to the user power utilization pattern, obtains corresponding using power mode
Feature.Data Dimensionality Reduction unit center determination unit.
The third aspect, the embodiment of the present invention also provide a kind of computer readable storage medium, and calculating is stored on the medium
Machine program can realize the above method when processor executes the computer program.
(3) advantageous effect
An embodiment of the present invention provides a kind of electricity consumption mode identification method of citizen requirement side and system, storage medium, tools
Standby following advantageous effect：
1, the embodiment of the present invention carries out dimension-reduction treatment using the first data set of stage feeding polymerization approximate data pair, obtains the second number
It, can be with the large-scale power load data of efficient process due to having carried out dimension-reduction treatment to the first data set according to collection；This hair
Bright embodiment determines initial cluster center, and then according to identified initial cluster center, using cluster according to the second data set
Algorithm is clustered, to identify user power utilization pattern.Due to determining initial cluster center according to the data set after dimensionality reduction, and
It is not using the initial cluster center randomly selected, and then raising cluster efficiency, that is, recognition efficiency；
2, the embodiment of the present invention uses the clustering algorithm based on dynamic time warping to third data set, so as to from difference
Similar residential electricity consumption pattern is identified in the residential electricity consumption load data curve of shape.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
Fig. 1 shows the flow diagram of the electricity consumption mode identification method of citizen requirement side in the embodiment of the present invention；
Fig. 2 shows the structure diagrams of the electricity consumption pattern recognition system of citizen requirement side in the embodiment of the present invention.
Specific implementation mode
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
The every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
The embodiment of the present invention provides a kind of electricity consumption mode identification method of citizen requirement side, as shown in Figure 1, this method packet
It includes：
S100, the power load data obtained in default historical time section；
In practical applications, intelligent electric meter can be set per the power load number for acquiring a user at regular intervals
According to for example, intelligent electric meter acquired a power load data every 30 minutes, such intelligent electric meter can be at 48 in one day
Power load data are acquired on time point, that is, acquire 48 power load data.For another example intelligent electric meter was acquired every 15 minutes
Power load data, such intelligent electric meter can acquire 96 power load data within one day.Further, it is possible to will
The power load data of intelligent electric meter acquisition pass through according to Customs Assigned Number in data transmission channel real-time Transmission to distributed data
The heart is stored.Wherein, each data center in distributive data center is arranged according to the location distribution of user, will use
Electric load data are stored to the data center in user region.
For example, when acquiring power load data, intelligent electric meter can acquire 96 power load data for one day, adopt
For the user's number integrated as m, it is n days to preset historical time section, the power load data of collected each user dimension
For 96n.
Certainly, after getting power load data, data cleansing can also be carried out to power load data, to check number
According to consistency, removal invalid value, missing values and duplicate data etc..
S200, the first data set is generated according to the power load data；Wherein, the line number of first data set and institute
State that the corresponding user's number of power load data is identical, the i-th row includes that i-th of user is each in the default historical time section
A power load data；
For convenience of description, the first data set can be denoted as to data set A, data set A can be indicated with following form：
User's number is m, has m row power load data, intelligent electric meter that can acquire within one day 96 use in data set A
Electric load data, it is n days to preset historical time section, the dimensions of the power load data of collected each user be 96n, institute
To there is 96n power load data in each row of data.For example, the first row power load data in data set A are first use
Each power load data of the family at n days.Wherein, a line power load data can form a load curve, so data
Collection A can form m load curve.
S300, each row power load data in first data set are carried out respectively using stage feeding polymerization approximate data
Dimension-reduction treatment obtains the second data set；
Wherein, stage feeding polymerization approximate data, i.e. Piecewise Aggregate Approximation, can be referred to as
PAA algorithms.
May be used in the specific implementation the first data set of various ways pair carry out dimensionality reduction to obtain the second data set, below
Introduce a kind of optional mode：
S301, each time interval in default historical time section is divided into multiple sub-districts according to user's living habit
Between；Wherein, the default historical time section includes multiple time intervals；
S302, following steps are executed per a line power load data in first data set：It calculates in the row
The mean value of each power load data in each time interval in each subinterval, and using the mean value as the subinterval
Corresponding power load data realize the dimension-reduction treatment to the row power load data.
For example, continuous n days that historical time section is nearest are preset, a time interval can be denoted as every day,
That is default historical time section includes n time interval, each time interval includes 96 power load data acquisitions
Time point.It was divided into r subinterval d i-th dayi=(t1,t2,…tr) (1≤i≤n), the length in each subinterval should be 15
Minute multiple, the time interval between two subintervals can use [tu,tu+1] (u ∈ [1, r-1]) expression.Then [t is usedu,
tu+1] in each power load data mean value replace [tu,tu+1] in all time points power load data.Here, sub-district
Between division can be divided according to user's living habit, for example, 24 can be divided by one day:00-6:00、6:00-9:00、9:
00-11:00、11:00-13:00、13:00-18:00、18:00-24:Calculate 6 mean values are made in 00 this 6 subintervals
Power load data for one day power load data, user's script each in this way are reduced to 6n by 96n dimensions.
S400, according to second data set, determine initial cluster center；
In the specific implementation, there are many modes for the determination of initial cluster center, for example, can be divided into user multigroup
And then determine initial cluster center, following manner specifically may be used and determine initial cluster center：
S401, according to preset Cluster Validity Index, determine the first clusters number；
There are many Cluster Validity Indexes, for example, Davies-Bouldin index (abbreviation DBI), SI etc..Assuming that using
DBI may be used following formula and determine the first clusters number as preset Cluster Validity Index：
Wherein, K ' is the first clusters number,Be in the i-th class each point to i classes center average distance,It is in jth class
Each point is to the average distance at j classes center, di,jIt is distance of the i-th class center to jth class center.The smaller expression Clustering Effects of DBI are got over
It is good, so that it is determined that the first clusters number.
S402, total electricity consumption of each user in the default historical time section is calculated；
For example, calculating each comfortable n days recently total electricity consumptions of this m user, m total electricity consumption is shared.
S403, the total electricity consumption of each user is ranked up, obtains ranking results；
, can from high to low for example, be ranked up to this m total electricity consumption, sequence that can also be from low to high.
S404, according to the ranking results, user is equally divided into K ' groups, and by the power load data of each group user
Mean value is as the initial cluster center；Wherein, K ' is first clusters number.
For example, user is divided into K ' groups, the mean value of each group power load data is as initial cluster center, at the beginning of shared K ' is a
Beginning cluster centre.
S500, it is based on the initial cluster center, each row in second data set is used using default clustering algorithm
Electric load data are clustered, and user power utilization pattern is obtained；.
For example, the default clustering algorithm includes K-means clustering algorithms.Wherein K-means algorithms are simple and efficient, and are being located
Superior performance when managing extensive intelligent electric meter data.It is of course also possible to use other kinds of clustering method, for example, EM algorithms,
Fuzzy clustering, hierarchical clustering, spectral clustering or the clustering method etc. based on neural network.
Since a line power load data can form a load curve, using clustering algorithm to each row electricity consumption data into
The process of row cluster, the process actually clustered to a plurality of load curve are clustered if load curve is similar
Together, a kind of user power utilization pattern is formed.
Above step S100~S500 is to identify user power utilization pattern from an angle.
In practical application, due to the change of the daily electricity consumption behavior of user, for example breakfast hour is postponed, and electricity consumption can be all caused
The shape of the corresponding load curve of load data changes, although the totality of user is not changed with power mode,
If user power utilization pattern be identified using traditional distance algorithm (for example, Euclidean distance), user power utilization mould can be caused
There is deviation in the identification of formula, is introduced from another angle recognition user power utilization pattern below by step S600~S700：
S600, third data set is generated according to the power load data；Wherein, the line number of the third data set is institute
The product of the number for the time interval that the corresponding user's number of power load data includes with the default historical time section is stated,
Power load data of each user in each time interval generate a line power load in the third data set
Data, the default historical time section includes multiple time intervals；
For convenience of description, third data set can be indicated with data set B：
Wherein, a line in data set B indicates a user in intraday 96 power load data, data set B
Every a line power load data can form a load curve, i.e. a user can be in intraday power load data
A load curve is formed, user's number is m, and the number for the time interval for including in default historical time section is n, data set B
Line number be m*n, m*n load curve can be formed.
S700, using the clustering algorithm based on dynamic time warping to each row power load number in the third data set
According to being clustered, user power utilization pattern is obtained；
The above-mentioned clustering algorithm (Dynamic Time Wrapping, abbreviation DTW) based on dynamic time warping can pass through
The shape similarity of two time series datas is weighed in the operations such as stretching, therefore DTW can be as the distance degree in clustering algorithm
Amount, so as to identify similar user power utilization pattern from load curve of different shapes.
In the specific implementation, the initial cluster center in S400 can be referred to as the first cluster centre, so with below
Involved in cluster process to cluster centre distinguish.The clustering algorithm pair based on dynamic time warping is used in step S700
The process that each row power load data in the third data set are clustered may comprise steps of：
S701, the second cluster centre is determined；
It will be appreciated that the second cluster centre that can be determined at random in step s 701 is the initial poly- of following cluster process
Class center.
S702, it is clustered using first function, and after each cluster, second cluster centre is updated,
Until the second cluster centre no longer changes or clusters number and reaches predetermined threshold value, user power utilization pattern is obtained；Described
One function is：
Wherein, B is third data set, and center is cluster centre, and DTW () is dynamic time warping function, and K is second poly-
Class number, k are more than or equal to 1 and are less than or equal to K, CkIndicate the power load data in kth class.
Above-mentioned first function can be referred to as the object function of cluster.Any two row vector a in data set BxAnd ayDynamically
Distance after Time alignment can be expressed as DTW (ax,ay)。
Above-mentioned predetermined threshold value, can according to circumstances sets itself, for example, 100 times.
Wherein, the determination process of the second clusters number may include：
DdR curves are drawn using following formula, and corresponding clusters number at the DdR knee of curves is gathered as described second
Class number；
In formula, D is first function, centeriFor ith cluster center, centerjFor j-th of cluster centre.
Wherein it is possible to be updated to second cluster centre using following formula：
In formula, center* is updated cluster centre, centerkIt is k-th of cluster centre.
S600~S700 through the above steps, the clustering algorithm based on dynamic time warping weigh the distance of similitude, into
And similar or similar user power utilization pattern can be identified from load curve of different shapes.
S800, feature extraction is carried out to user power utilization pattern, obtains corresponding electricity consumption pattern feature.
Above-mentioned user power utilization pattern actually indicates with the form of curve, dissimilar load curve cluster to be formed it is different
User power utilization pattern, so-called feature extraction are actually to be expressed by the way of verbal description to user power utilization pattern,
That is electricity consumption pattern feature, so as to people's more intuitive understanding user power utilization pattern.
The embodiment of the present invention provides a kind of citizen requirement side electricity consumption mode identification method, uses stage feeding polymerization approximation to calculate first
The first data set of method pair carries out dimension-reduction treatment, obtains the second data set, and then according to the second data set, determine in initial clustering
The heart, and then according to identified initial cluster center, clustered using clustering algorithm, to identify user power utilization pattern.
The embodiment of the present invention, can be with the large-scale power load number of efficient process due to having carried out dimension-reduction treatment to the first data set
According to, and due to determining initial cluster center according to the data set after dimensionality reduction, rather than use in the initial clustering randomly selected
The heart, and then improve cluster efficiency.The embodiment of the present invention uses the clustering algorithm based on dynamic time warping to third data set, from
And similar residential electricity consumption pattern can be identified from residential electricity consumption load data curve of different shapes.
The embodiment of the present invention also provides a kind of electricity consumption pattern recognition system of citizen requirement side, the system and above-mentioned electricity consumption mould
Formula recognition methods is corresponding, as shown in Fig. 2, the system include data acquisition module, the first identification module or the second identification module,
And characteristic extracting module, wherein：
The data acquisition module is used to obtain the power load data in default historical time section；First identification
Module includes the first generation unit, Data Dimensionality Reduction unit, center determination unit and the first cluster cell；First generation unit
For generating the first data set according to the power load data；Wherein, the line number of first data set and the electricity consumption are negative
The corresponding user's number of lotus data is identical, and the i-th row includes that each electricity consumption of i-th of user in the default historical time section is negative
Lotus data；The Data Dimensionality Reduction unit is used for negative to each row electricity consumption in first data set using stage feeding polymerization approximate data
Lotus data carry out dimension-reduction treatment respectively, obtain the second data set；The center determination unit is used for according to second data set,
Determine initial cluster center；First cluster cell is used to be based on the initial cluster center, using default clustering algorithm pair
Each row power load data in second data set are clustered, and user power utilization pattern is obtained；
Second identification module includes the second generation unit and the second cluster cell；Second generation unit is used for according to institute
It states power load data and generates third data set；Wherein, the line number of the third data set corresponds to for the power load data
User's number and the default historical time section time interval that includes number product, each user is at each
Power load data in time interval generate a line power load data in the third data set, when the default history
Between section include multiple time intervals；Second cluster cell is used for using the clustering algorithm based on dynamic time warping to institute
Each row power load data stated in third data set are clustered, and user power utilization pattern is obtained；
The characteristic extracting module is used to carry out feature extraction to the user power utilization pattern, obtains corresponding using power mode
Feature.
In some embodiments, Data Dimensionality Reduction unit is specifically used for：By each time zone in default historical time section
Between according to user's living habit be divided into multiple subintervals；Wherein, the default historical time section includes multiple time intervals；Needle
To executing following steps per a line power load data in first data set：It calculates in each time interval in the row
The mean value of each power load data in each subinterval, and using the mean value as the corresponding power load number in the subinterval
According to dimension-reduction treatment of the realization to the row power load data.
In some embodiments, center determination unit is specifically used for：According to preset Cluster Validity Index, first is determined
Clusters number；Calculate total electricity consumption of each user in the default historical time section；To the total electricity consumption of each user into
Row sequence, obtains ranking results；According to the ranking results, user is equally divided into k ' groups, and the electricity consumption of each group user is born
The mean value of lotus data is as the initial cluster center；Wherein, k ' is first clusters number.
In some embodiments, the default clustering algorithm includes K-means clustering algorithms.
In some embodiments, the initial cluster center is the first cluster centre；Second cluster cell is specifically used
In：It is random to determine the second cluster centre；It is clustered using first function, and after each cluster, in second cluster
The heart is updated, until the second cluster centre no longer changes or clusters number and reaches predetermined threshold value, obtains user power utilization
Pattern；The first function is：
Wherein, B is third data set, and center is cluster centre, and DTW () is dynamic time warping function, and K is second poly-
Class number, k are more than or equal to 1 and are less than or equal to K, CkIndicate the power load data in kth class.
In some embodiments, the second cluster cell determines that the process of second clusters number includes：It is painted using following formula
DdR curves processed, and using corresponding clusters number at the DdR knee of curves as second clusters number；
Wherein, D is first function, centeriFor ith cluster center, centerjFor j-th of cluster centre.
In some embodiments, the second cluster cell is updated second cluster centre using following formula：
Wherein, center* is updated cluster centre, centerkIt is k-th of cluster centre.
It will be appreciated that electricity consumption pattern recognition system provided in an embodiment of the present invention is opposite with electricity consumption mode identification method
It answers, the contents such as explanation, citing, embodiment, advantageous effect in relation to content can refer in above-mentioned electricity consumption mode identification method
Corresponding contents, do not repeating herein.
In addition, the embodiment of the present invention also provides a kind of computer readable storage medium, computer journey is stored on the medium
Sequence can realize the above method when processor executes the computer program.
It should be noted that herein, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also include other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although with reference to the foregoing embodiments
Invention is explained in detail, it will be understood by those of ordinary skill in the art that：It still can be to aforementioned each implementation
Technical solution recorded in example is modified or equivalent replacement of some of the technical features；And these modification or
It replaces, the spirit and scope for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (9)
1. a kind of electricity consumption mode identification method of citizen requirement side, which is characterized in that including：
Obtain the power load data in default historical time section；
The first data set is generated according to the power load data；Wherein, the line number of first data set and the electricity consumption are negative
The corresponding user's number of lotus data is identical, and the i-th row includes that each electricity consumption of i-th of user in the default historical time section is negative
Lotus data；Each row power load data in first data set are carried out at dimensionality reduction respectively using stage feeding polymerization approximate data
Reason, obtains the second data set；According to second data set, initial cluster center is determined；Based on the initial cluster center, adopt
Each row power load data in second data set are clustered with default clustering algorithm, obtain user power utilization pattern；
Alternatively, generating third data set according to the power load data；Wherein, the line number of the third data set is the use
The product of the number for the time interval that the corresponding user's number of electric load data includes with the default historical time section, it is each
Power load data of a user in each time interval generate a line power load data in the third data set,
The default historical time section includes multiple time intervals；Using the clustering algorithm based on dynamic time warping to the third
Each row power load data in data set are clustered, and user power utilization pattern is obtained；
Feature extraction is carried out to the user power utilization pattern, obtains corresponding electricity consumption pattern feature.
2. according to the method described in claim 1, it is characterized in that, described counted using stage feeding polymerization approximate data to described first
Dimension-reduction treatment is carried out respectively according to each row power load data of concentration, including：
Each time interval in default historical time section is divided into multiple subintervals according to user's living habit；Wherein,
The default historical time section includes multiple time intervals；
For in first data set following steps are executed per a line power load data：
The mean value of each power load data in each time interval in the row in each subinterval is calculated, and should
Mean value realizes the dimension-reduction treatment to the row power load data as the corresponding power load data in the subinterval.
3. the method as described in claim 1, which is characterized in that it is described according to second data set, it determines in initial clustering
The heart, including：
According to preset Cluster Validity Index, the first clusters number is determined；
Calculate total electricity consumption of each user in the default historical time section；
The total electricity consumption of each user is ranked up, ranking results are obtained；
According to the ranking results, user is equally divided into k ' groups, and using the mean value of the power load data of each group user as
The initial cluster center；Wherein, k ' is first clusters number.
4. according to the method described in claim 1, it is characterized in that, the default clustering algorithm includes K-means clustering algorithms.
5. according to the method described in claim 1, it is characterized in that, the initial cluster center is the first cluster centre；It is described
Each row power load data in the third data set are clustered using the clustering algorithm based on dynamic time warping, are wrapped
It includes：
Determine the second cluster centre；
It is clustered using first function, and after each cluster, to second cluster centre
It is updated, until the second cluster centre no longer changes or clusters number and reaches predetermined threshold value, obtains user's use
Power mode；The first function is：
Wherein, B is third data set, and center is cluster centre, and DTW () is dynamic time warping function, and K is the second cluster numbers
Mesh, k are more than or equal to 1 and are less than or equal to K, CkIndicate the power load data in kth class.
6. according to the method described in claim 5, it is characterized in that, the determination process of second clusters number includes：
DdR curves are drawn using following formula, and using corresponding clusters number at the DdR knee of curves as second cluster numbers
Mesh；
Wherein, D is first function, centeriFor ith cluster center, centerjFor j-th of cluster centre.
7. according to the method described in claim 5, it is characterized in that, being updated to second cluster centre using following formula：
Wherein, center* is updated cluster centre, centerkIt is k-th of cluster centre.
8. a kind of electricity consumption pattern recognition system, which is characterized in that including data acquisition module, the first identification module or the second identification
Module and characteristic extracting module, wherein：
The data acquisition module is used to obtain the power load data in default historical time section；First identification module
Including the first generation unit, Data Dimensionality Reduction unit, center determination unit and the first cluster cell；First generation unit is used for
The first data set is generated according to the power load data；Wherein, the line number of first data set and the power load number
Identical according to corresponding user's number, the i-th row includes each power load number of i-th of user in the default historical time section
According to；The Data Dimensionality Reduction unit is used for using stage feeding polymerization approximate data to each row power load number in first data set
According to dimension-reduction treatment is carried out respectively, the second data set is obtained；The center determination unit is used to, according to second data set, determine
Initial cluster center；First cluster cell is used to be based on the initial cluster center, using default clustering algorithm to described
Each row power load data in second data set are clustered, and user power utilization pattern is obtained；
Second identification module includes the second generation unit and the second cluster cell；Second generation unit is used for according to the use
Electric load data generate third data set；Wherein, the line number of the third data set is the corresponding use of the power load data
The product of the number for the time interval that family number includes with the default historical time section, each user is in each time
Power load data in section generate a line power load data in the third data set, the default historical time section
Include multiple time intervals；Second cluster cell is used for using the clustering algorithm based on dynamic time warping to described the
Each row power load data in three data sets are clustered, and user power utilization pattern is obtained；
The characteristic extracting module is used to carry out feature extraction to the user power utilization pattern, obtains corresponding special with power mode
Sign.
9. a kind of computer readable storage medium, computer program is stored on the medium, which is characterized in that execute in processor
Claim 1~7 any method can be realized when the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810130203.0A CN108460410A (en) | 2018-02-08 | 2018-02-08 | Electricity consumption mode identification method and system, the storage medium of citizen requirement side |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810130203.0A CN108460410A (en) | 2018-02-08 | 2018-02-08 | Electricity consumption mode identification method and system, the storage medium of citizen requirement side |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108460410A true CN108460410A (en) | 2018-08-28 |
Family
ID=63239018
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810130203.0A Pending CN108460410A (en) | 2018-02-08 | 2018-02-08 | Electricity consumption mode identification method and system, the storage medium of citizen requirement side |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108460410A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109493249A (en) * | 2018-11-05 | 2019-03-19 | 北京邮电大学 | A kind of analysis method of electricity consumption data on Multiple Time Scales |
| CN109492048A (en) * | 2019-01-21 | 2019-03-19 | 国网河北省电力有限公司经济技术研究院 | A kind of extracting method, system and the terminal device of power consumer electrical characteristics |
| CN109766907A (en) * | 2018-11-23 | 2019-05-17 | 国网江苏省电力有限公司电力科学研究院 | A kind of trade power consumption schema extraction method for supporting pattern cycle self-discovery |
| CN109783528A (en) * | 2018-11-23 | 2019-05-21 | 国网江苏省电力有限公司电力科学研究院 | A kind of electricity consumption schema extraction method and system |
| CN109934301A (en) * | 2019-03-22 | 2019-06-25 | 广东电网有限责任公司 | A kind of power load aggregation analysis method, device and equipment |
| CN110826641A (en) * | 2019-11-13 | 2020-02-21 | 上海积成能源科技有限公司 | System and method for classifying electricity consumption condition of residents based on cluster analysis |
| CN111199016A (en) * | 2019-09-29 | 2020-05-26 | 国网湖南省电力有限公司 | DTW-based improved K-means daily load curve clustering method |
| CN111612650A (en) * | 2020-05-27 | 2020-09-01 | 福州大学 | Power consumer clustering method and system based on DTW distance and neighbor propagation clustering algorithm |
| CN113449793A (en) * | 2021-06-28 | 2021-09-28 | 国网北京市电力公司 | Method and device for determining power utilization state |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015059272A1 (en) * | 2013-10-24 | 2015-04-30 | Universite Libre De Bruxelles | Improved non-intrusive appliance load monitoring method and device |
| CN105004939A (en) * | 2015-05-08 | 2015-10-28 | 浙江工业大学 | Composite electric energy quality disturbance signal quantitative analysis method |
| CN106022959A (en) * | 2016-05-19 | 2016-10-12 | 北京中电普华信息技术有限公司 | Peak clipping and valley filling-oriented electricity utilization behavior analysis method and system |
| CN106600119A (en) * | 2016-11-30 | 2017-04-26 | 南方电网科学研究院有限责任公司 | K-means-based power consumer clustering method and device |
| CN107657266A (en) * | 2017-08-03 | 2018-02-02 | 华北电力大学（保定） | A kind of load curve clustering method based on improvement spectrum multiple manifold cluster |
-
2018
- 2018-02-08 CN CN201810130203.0A patent/CN108460410A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015059272A1 (en) * | 2013-10-24 | 2015-04-30 | Universite Libre De Bruxelles | Improved non-intrusive appliance load monitoring method and device |
| CN105004939A (en) * | 2015-05-08 | 2015-10-28 | 浙江工业大学 | Composite electric energy quality disturbance signal quantitative analysis method |
| CN106022959A (en) * | 2016-05-19 | 2016-10-12 | 北京中电普华信息技术有限公司 | Peak clipping and valley filling-oriented electricity utilization behavior analysis method and system |
| CN106600119A (en) * | 2016-11-30 | 2017-04-26 | 南方电网科学研究院有限责任公司 | K-means-based power consumer clustering method and device |
| CN107657266A (en) * | 2017-08-03 | 2018-02-02 | 华北电力大学（保定） | A kind of load curve clustering method based on improvement spectrum multiple manifold cluster |
Non-Patent Citations (3)
| Title |
|---|
| THANCHANOK TEERARATKUL等: "shape-based approach to household electric load curve clustering and prediction", 《IEEE TRANSACTIONS ON SMART GRID》 * |
| 林顺富等: "基于信息熵分段聚合近似和谱聚类的负荷分类方法", 《中国电机工程学报》 * |
| 沈建良等: "大数据技术在电力大用户用电特征分类中的应用", 《浙江电力》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109493249B (en) * | 2018-11-05 | 2021-11-12 | 北京邮电大学 | Analysis method of electricity consumption data on multiple time scales |
| CN109493249A (en) * | 2018-11-05 | 2019-03-19 | 北京邮电大学 | A kind of analysis method of electricity consumption data on Multiple Time Scales |
| CN109766907A (en) * | 2018-11-23 | 2019-05-17 | 国网江苏省电力有限公司电力科学研究院 | A kind of trade power consumption schema extraction method for supporting pattern cycle self-discovery |
| CN109783528A (en) * | 2018-11-23 | 2019-05-21 | 国网江苏省电力有限公司电力科学研究院 | A kind of electricity consumption schema extraction method and system |
| CN109783528B (en) * | 2018-11-23 | 2019-09-10 | 国网江苏省电力有限公司电力科学研究院 | A kind of electricity consumption schema extraction method and system |
| CN109492048A (en) * | 2019-01-21 | 2019-03-19 | 国网河北省电力有限公司经济技术研究院 | A kind of extracting method, system and the terminal device of power consumer electrical characteristics |
| CN109934301A (en) * | 2019-03-22 | 2019-06-25 | 广东电网有限责任公司 | A kind of power load aggregation analysis method, device and equipment |
| CN109934301B (en) * | 2019-03-22 | 2022-10-04 | 广东电网有限责任公司 | Power load cluster analysis method, device and equipment |
| CN111199016A (en) * | 2019-09-29 | 2020-05-26 | 国网湖南省电力有限公司 | DTW-based improved K-means daily load curve clustering method |
| CN111199016B (en) * | 2019-09-29 | 2023-11-07 | 国网湖南省电力有限公司 | Daily load curve clustering method for improving K-means based on DTW |
| CN110826641A (en) * | 2019-11-13 | 2020-02-21 | 上海积成能源科技有限公司 | System and method for classifying electricity consumption condition of residents based on cluster analysis |
| CN111612650B (en) * | 2020-05-27 | 2022-06-17 | 福州大学 | DTW distance-based power consumer grouping method and system |
| CN111612650A (en) * | 2020-05-27 | 2020-09-01 | 福州大学 | Power consumer clustering method and system based on DTW distance and neighbor propagation clustering algorithm |
| CN113449793A (en) * | 2021-06-28 | 2021-09-28 | 国网北京市电力公司 | Method and device for determining power utilization state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108460410A (en) | Electricity consumption mode identification method and system, the storage medium of citizen requirement side | |
| CN106410781B (en) | A kind of power consumer demand response Potential Determining Method | |
| CN108376262B (en) | Analytical model construction method for typical characteristics of wind power output | |
| CN106815677A (en) | The recognition methods of non-intrusion type load and device | |
| CN106408008A (en) | Load curve distance and shape-based load classification method | |
| CN107133652A (en) | Electricity customers Valuation Method and system based on K means clustering algorithms | |
| CN106203478A (en) | A kind of load curve clustering method for the big data of intelligent electric meter | |
| CN102004962A (en) | Method for realizing intelligent collation on personal comprehensive performance for appraisal | |
| CN109978404A (en) | Power transmission channel expansion planning method considering uncertainty factors and scene reduction | |
| CN112053059A (en) | Peak-valley load optimization method based on user demand elastic response and time-of-use electricity price | |
| CN109117872A (en) | A kind of user power utilization behavior analysis method based on automatic Optimal Clustering | |
| CN105844294A (en) | Electricity usage behavior analysis method based on FCM cluster algorithm | |
| CN109634940A (en) | A kind of typical low pressure platform area's electricity consumption model building method based on magnanimity low-voltage platform area electricity consumption data | |
| CN110109971A (en) | A kind of low-voltage platform area user power utilization Load Characteristic Analysis method | |
| CN106528804B (en) | A kind of tenant group method based on fuzzy clustering | |
| CN108765194A (en) | A kind of effective residential electricity consumption behavior analysis system | |
| CN110298552A (en) | A kind of power distribution network individual power method for detecting abnormality of combination history electrical feature | |
| CN107292489A (en) | A kind of dispatching of power netwoks runs lean evaluation system | |
| CN106022578A (en) | Residential electricity peak-valley-flat dividing method based on data dimension increasing and K-means clustering | |
| Wang et al. | Aggregated large-scale air-conditioning load: Modeling and response capability evaluation of virtual generator units | |
| CN111914900A (en) | User power consumption mode classification method | |
| CN107918779A (en) | One kind builds polynary load characteristics clustering model method and system | |
| Cheong et al. | Unveiling the relationship between economic growth and equality for developing countries | |
| CN106354886A (en) | Method for screening nearest neighbor by using potential neighbor relation graph in recommendation system | |
| CN111695599B (en) | Elastic identification method for user electricity load time |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180828 |