CN109284851A - A kind of user power utilization behavior classification method suitable for Demand Side Response - Google Patents
A kind of user power utilization behavior classification method suitable for Demand Side Response Download PDFInfo
- Publication number
- CN109284851A CN109284851A CN201810594612.6A CN201810594612A CN109284851A CN 109284851 A CN109284851 A CN 109284851A CN 201810594612 A CN201810594612 A CN 201810594612A CN 109284851 A CN109284851 A CN 109284851A
- Authority
- CN
- China
- Prior art keywords
- power utilization
- cluster
- user power
- demand side
- side response
- 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
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- 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 invention discloses a kind of user power utilization behavior classification methods suitable for Demand Side Response, pass through third index flatness polishing AFR control, load data normalization first；Then user power utilization load curve is clustered using the K-means algorithm of evolution, determines the type of residential electricity consumption behavior；The electricity consumption behavior to Add User is divided into known load curve cluster by the update that finally the electricity consumption behavior to Add User is classified with improved support vector machines method.The present invention has good generalization ability, and the user power utilization behavioral trait that can be suitable for various industries is classified.
Description
Technical field
The invention belongs to electric system Demand-side technical field of information management, and in particular to one kind is suitable for Demand Side Response
User power utilization behavior classification method.
Background technique
With the all-round construction of smart grid, Information and Communication Technology and smart grid are closely connected together, while sale of electricity
Side is decontroled, and the advanced measuring system based on intelligent electric meter can extend to ordinary user, the intelligence of intelligent power field
It is gradually increased with interactive level.The propulsion of power grid reform is abundant to user power utilization behavior so that the status of user significantly improves
Understanding facilitates Utilities Electric Co. and promotes competitiveness when fighting for user resources, and is also the basis that Demand Side Response is achieved
And premise.Demand response is development of the demand side management in Competitive Electricity Market, by price signal and incentive mechanism come
It increases demand the effect of side in the market, can effectively alleviate transmission of electricity and paces that the capacity that generates electricity is extended rapidly, be conducive to structure
Build friendly environment society.Therefore, the electricity consumption property sort technique study interacted suitable for user with power grid close friend is for intelligence
The development trend of energy electricity consumption, the effective user power utilization property sort method of proposition reduce peak-valley difference for Demand Side Response, avoid the peak hour
Management, electricity pricing and load prediction provide effective support.
Summary of the invention
In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is that providing a kind of suitable for needing
The user power utilization behavior classification method for asking side to respond instructs the electricity consumption behavior of user's progress efficiently and conveniently for Utilities Electric Co..
The invention adopts the following technical scheme:
A kind of user power utilization behavior classification method suitable for Demand Side Response passes through third index flatness polishing first
AFR control, load data normalization；Then user power utilization load curve is clustered using the K-means algorithm of evolution, is determined
The type of residential electricity consumption behavior out；Finally the electricity consumption behavior to Add User is classified more with improved support vector machines method
Newly, the electricity consumption behavior to Add User is divided into known load curve cluster.
Specifically, third index flatness polishing AFR control specifically: set daily load sequence as X1,...,XT, so
Data normalization processing is carried out using the prediction model of third index flatness afterwards, load data is restricted in setting range.
Further, the prediction model of third index flatnessIt is specific as follows:
Wherein, m=1,2 ..., StFor intermediate variable, β is sequential parameter.
Further, intermediate variable S three timestSpecific calculating is as follows:
。
Specifically, improve K-means algorithm using hierarchical clustering method, first application level clustering algorithm obtain one it is initial
It divides, calculates the mean value of object in each class, and using it as the initial cluster center of K-mean algorithm.
Further, the type of residential electricity consumption behavior is determined using the K-means algorithm of evolution specifically: when use K-
Load curve has been divided into K cluster by means algorithm, for each vector in cluster, calculates separately their silhouette coefficient, profile
The value of coefficient selects silhouette coefficient to cluster close to 1 classification number as K-means between [- 1,1] in the cluster of load curve
K value.
Further, the silhouette coefficient of all the points is averaging, obtains the total silhouette coefficient of the cluster result, profile system
A point i vector silhouette coefficient in number is as follows:
Wherein, a (i) is the average value that other put dissimilar degree in i vector to same cluster, and b (i) is that i vector arrives other
The minimum value of the average dissimilar degree of cluster.
Specifically, improved support vector machines method specifically: test different kernel functions respectively, after determining kernel function, give
Determine parameter C and the biggish value interval of parameter γ range, grid section is marked off by parameter C and parameter γ, then use is enumerated
Method finds optimal parameter C and parameter γ and obtains classification results.
Compared with prior art, the present invention at least has the advantages that
Algorithm proposed by the present invention third index flatness polishing missing data, the complete of data are conducive to improve cluster
Then user power utilization load curve is clustered using the K-means algorithm of evolution, determines the kind of residential electricity consumption behavior by effect
Class；The update that finally the electricity consumption behavior to Add User is classified with improved support vector machines method, the use to Add User
Electric behavior is divided into known load curve cluster, after the label for determining user, the improved branch of electricity consumption behavior that Adds User
It holds vector machine algorithm to classify, the classification that will Add User of efficient quick.
Further, it determines that cluster classification, hierarchical clustering method determine cluster centre with silhouette coefficient method, reduces iteration time
Number, improves arithmetic speed, improves the accuracy of cluster.
Further, after the label for determining user, the improved algorithm of support vector machine of electricity consumption behavior that Adds User
Classify, the classification that will Add User of efficient quick.
In conclusion the present invention has good generalization ability, the user power utilization behavior that can be suitable for various industries is special
Property classification.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Detailed description of the invention
Fig. 1 is user power utilization characteristic behavior classification process figure.
Specific embodiment
Referring to Fig. 1, a kind of user power utilization behavior classification method suitable for Demand Side Response of the present invention, provides three first
Secondary exponential smoothing polishing AFR control；Then user power utilization load curve is clustered with the K-means algorithm evolved, is determined
The type of residential electricity consumption behavior；Finally the electricity consumption behavior to Add User is classified more with improved support vector machines method
Newly, the electricity consumption behavior to Add User is divided into known load curve cluster.
Wherein, the K-means algorithm of evolution silhouette coefficient method determines that cluster classification, hierarchical clustering method determine in cluster
The heart.
In support vector cassification, in order to can achieve better classifying quality, different kernel functions is tested respectively.Really
After fixed a certain specific kernel function, given parameters C and the biggish value interval of parameter γ range thus can be by parameters
C and parameter γ mark off grid section, and optimal parameter C and parameter γ are then found with the method enumerated, it is higher to obtain precision
Classification results.
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.The present invention being described and shown in usually here in attached drawing is real
The component for applying example can be arranged and be designed by a variety of different configurations.Therefore, below to the present invention provided in the accompanying drawings
The detailed description of embodiment be not intended to limit the range of claimed invention, but be merely representative of of the invention selected
Embodiment.Based on the embodiments of the present invention, those of ordinary skill in the art are obtained without creative efforts
The every other embodiment obtained, shall fall within the protection scope of the present invention.
Referring to Fig. 1, the present invention provides a kind of user power utilization behavior classification method suitable for Demand Side Response, using three times
Exponential smoothing polishing AFR control is clustered with the K-means algorithm of evolution and determines user tag, the electricity consumption behavior to Add User
Sorted out with improved algorithm of support vector machine, analyze the use electrical characteristics of different set of curves, Utilities Electric Co. can efficiently, just
Consumer-oriented electricity consumption behavior promptly.
1) data prediction；
1.1) AFR control uses third index flatness polishing, if daily load sequence is X1,...,XT, calculate
Formula is
The prediction model of third index flatness is
Wherein, m=1,2 ...,
1.2) data normalization is handled, and load data is restricted in a certain range.
2) the K-means algorithm cluster evolved；
2.1) silhouette coefficient method determines cluster classification
When load curve has been divided into K cluster using K-means algorithm.For each vector in cluster, it is calculated separately
Silhouette coefficient.The value of silhouette coefficient is between [- 1,1], and more leveling off to 1, to represent cohesion degree and separating degree all relatively excellent.It is right
For one of point i, i vector silhouette coefficient is just are as follows:
Wherein a (i) is the average value that other put dissimilar degree in i vector to same cluster, and b (i) is i vector to other clusters
Average dissimilar degree minimum value, the silhouette coefficient of all the points is averaging, is exactly the total silhouette coefficient of the cluster result.
The K value for selecting silhouette coefficient to cluster closer to the classification number for 1 as K-means in the cluster of load curve；
2.2) hierarchical clustering method determines cluster centre
Hierarchy clustering method is exactly by carrying out hierachical decomposition according to some way to data set, until meeting certain condition
Until.For K-means algorithm to initial cluster center initial value sensitive issue, K- can be improved using hierarchical clustering method
Means algorithm.First application level clustering algorithm obtains an initial division, calculates the mean value of object in each class, and by it
Initial cluster center as K-mean algorithm.
3) improved algorithm of support vector machine
In support vector cassification, the selection of punishment parameter C be it is uncertain, its value will affect classification results.
It is also not identical using the classifying quality of different kernel functions when the data characteristics difference of training set.For Polynomial kernel function,
Radial basis kernel function and S-shaped kernel function can also be related to the setting of γ parameter.
In order to can achieve better classifying quality, different kernel functions is tested respectively.Determine a certain specific core letter
After number, given parameters C and the biggish value interval of parameter γ range thus can mark off net by parameter C and parameter γ
Then lattice section finds optimal parameter C and parameter γ with the method enumerated, obtains the higher classification results of precision.
The above content is merely illustrative of the invention's technical idea, and this does not limit the scope of protection of the present invention, all to press
According to technical idea proposed by the present invention, any changes made on the basis of the technical scheme each falls within claims of the present invention
Protection scope within.
Claims (8)
1. a kind of user power utilization behavior classification method suitable for Demand Side Response, which is characterized in that pass through index three times first
Exponential smoothing polishing AFR control, load data normalization；Then utilize the K-means algorithm of evolution by user power utilization load curve
Cluster, determines the type of residential electricity consumption behavior；Finally by the electricity consumption behavior to Add User with improved support vector machines method into
The electricity consumption behavior to Add User is divided into known load curve cluster by the update of row classification.
2. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 1, feature exist
In third index flatness polishing AFR control specifically: set daily load sequence as X1,...,XT, then using referring to three times
The prediction model of number exponential smoothing carries out data normalization processing, and load data is restricted in setting range.
3. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 2, feature exist
In the prediction model of third index flatnessIt is specific as follows:
Wherein, m=1,2 ..., StFor intermediate variable, β is sequential parameter.
4. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 3, feature exist
In intermediate variable S three timestSpecific calculating is as follows:
。
5. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 1, feature exist
In using hierarchical clustering method improvement K-means algorithm, first application level clustering algorithm obtains an initial division, calculates every
The mean value of object in a class, and using it as the initial cluster center of K-mean algorithm.
6. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 5, feature exist
In determining the type of residential electricity consumption behavior using the K-means algorithm of evolution specifically: will be born when using K-means algorithm
Lotus curve has been divided into K cluster, for each vector in cluster, calculates separately their silhouette coefficient, the value of silhouette coefficient between
[- 1,1], the K value for selecting classification number of the silhouette coefficient close to 1 to cluster in the cluster of load curve as K-means.
7. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 6, feature exist
In, the silhouette coefficients of all the points is averaging, the total silhouette coefficient of the cluster result is obtained, a point i in silhouette coefficient to
It is as follows to measure silhouette coefficient:
Wherein, a (i) is the average value that other put dissimilar degree in i vector to same cluster, and b (i) is i vector to other clusters
The minimum value of average dissmilarity degree.
8. a kind of user power utilization behavior classification method suitable for Demand Side Response according to claim 1, feature exist
In improved support vector machines method specifically: different kernel functions is tested respectively, after determining kernel function, given parameters C and parameter
The biggish value interval of γ range marks off grid section by parameter C and parameter γ, is then found with the method enumerated optimal
Parameter C and parameter γ obtain classification results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810594612.6A CN109284851A (en) | 2018-06-11 | 2018-06-11 | A kind of user power utilization behavior classification method suitable for Demand Side Response |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810594612.6A CN109284851A (en) | 2018-06-11 | 2018-06-11 | A kind of user power utilization behavior classification method suitable for Demand Side Response |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109284851A true CN109284851A (en) | 2019-01-29 |
Family
ID=65186252
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810594612.6A Pending CN109284851A (en) | 2018-06-11 | 2018-06-11 | A kind of user power utilization behavior classification method suitable for Demand Side Response |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109284851A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781332A (en) * | 2019-10-16 | 2020-02-11 | 三峡大学 | Electric power resident user daily load curve clustering method based on composite clustering algorithm |
| CN110851502A (en) * | 2019-11-19 | 2020-02-28 | 国网吉林省电力有限公司 | Load characteristic scene classification method based on data mining technology |
| CN111047465A (en) * | 2019-12-09 | 2020-04-21 | 湘潭大学 | Power grid friendly load response method based on electric power big data |
| CN112560934A (en) * | 2020-12-10 | 2021-03-26 | 国家电网有限公司 | Multi-scale power consumption behavior analysis method |
| CN113435623A (en) * | 2020-03-23 | 2021-09-24 | 北京沃东天骏信息技术有限公司 | Early warning method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117810A (en) * | 2015-09-24 | 2015-12-02 | 国网福建省电力有限公司泉州供电公司 | Residential electricity consumption mid-term load prediction method under multistep electricity price mechanism |
| CN107480341A (en) * | 2017-07-21 | 2017-12-15 | 河海大学 | A kind of dam safety comprehensive method based on deep learning |
| CN107680391A (en) * | 2017-09-28 | 2018-02-09 | 长沙理工大学 | Two pattern fuzzy control methods of crossroad access stream |
-
2018
- 2018-06-11 CN CN201810594612.6A patent/CN109284851A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117810A (en) * | 2015-09-24 | 2015-12-02 | 国网福建省电力有限公司泉州供电公司 | Residential electricity consumption mid-term load prediction method under multistep electricity price mechanism |
| CN107480341A (en) * | 2017-07-21 | 2017-12-15 | 河海大学 | A kind of dam safety comprehensive method based on deep learning |
| CN107680391A (en) * | 2017-09-28 | 2018-02-09 | 长沙理工大学 | Two pattern fuzzy control methods of crossroad access stream |
Non-Patent Citations (4)
| Title |
|---|
| 周莹: ""基于证券理财产品用户行为分析的个性化推荐研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张靖 等: ""优化初始聚类中心的改进k-means算法"", 《计算机工程与设计》 * |
| 王梅: ""一种改进的核函数参数选择方法"", 《中国优秀硕士学位论文全文数据库 基础科学辑》 * |
| 黄艳红 等: ""基于三次指数平滑对用户用电量的预测"", 《上海工程技术大学学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781332A (en) * | 2019-10-16 | 2020-02-11 | 三峡大学 | Electric power resident user daily load curve clustering method based on composite clustering algorithm |
| CN110851502A (en) * | 2019-11-19 | 2020-02-28 | 国网吉林省电力有限公司 | Load characteristic scene classification method based on data mining technology |
| CN111047465A (en) * | 2019-12-09 | 2020-04-21 | 湘潭大学 | Power grid friendly load response method based on electric power big data |
| CN113435623A (en) * | 2020-03-23 | 2021-09-24 | 北京沃东天骏信息技术有限公司 | Early warning method and device |
| CN112560934A (en) * | 2020-12-10 | 2021-03-26 | 国家电网有限公司 | Multi-scale power consumption behavior analysis method |
| CN112560934B (en) * | 2020-12-10 | 2024-03-01 | 国家电网有限公司 | Multi-scale power consumption behavior analysis method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109284851A (en) | A kind of user power utilization behavior classification method suitable for Demand Side Response | |
| Yang et al. | A review of electric load classification in smart grid environment | |
| Huang et al. | A clustering based grouping method of nearly zero energy buildings for performance improvements | |
| CN108681973A (en) | Sorting technique, device, computer equipment and the storage medium of power consumer | |
| Kim et al. | Methods for generating TLPs (typical load profiles) for smart grid-based energy programs | |
| CN107423769A (en) | Electric load curve adaptive clustering scheme based on morphological feature | |
| CN107833153B (en) | Power grid load missing data completion method based on k-means clustering | |
| CN104200275A (en) | Power utilization mode classification and control method based on user behavior characteristics | |
| CN108694673A (en) | A kind of processing method, device and the processing equipment of insurance business risk profile | |
| CN108805193A (en) | A kind of power loss data filling method based on mixed strategy | |
| CN103679555A (en) | Time-of-use electricity price determining method based on load characteristic classification | |
| CN107944604A (en) | A kind of weather pattern recognition methods and device for photovoltaic power prediction | |
| CN109146252A (en) | Based on the user power utilization behavior clustering method and system for adjusting potentiality index | |
| CN108074004A (en) | A kind of GIS-Geographic Information System short-term load forecasting method based on gridding method | |
| CN112149873A (en) | Low-voltage transformer area line loss reasonable interval prediction method based on deep learning | |
| Yan | Research on green suppliers' evaluation based on AHP & genetic algorithm | |
| CN112232863B (en) | Virtual resource management method and system based on cloud edge-side cooperation | |
| CN110598968A (en) | Power grid investment benefit evaluation method based on improved matter element extension model | |
| CN109544029A (en) | Analysis method, analytical equipment and the terminal of a kind of area's line loss | |
| CN113902304A (en) | Controllable load screening method based on total load curve similarity | |
| CN115861671A (en) | Double-layer self-adaptive clustering method considering load characteristics and adjustable potential | |
| CN116148753A (en) | Intelligent electric energy meter operation error monitoring system | |
| CN105653670A (en) | Intelligent electricity consumption data mining method based on manifold learning clustering algorithm | |
| CN110321390A (en) | Based on the load curve data visualization method for thering is supervision and unsupervised algorithm to combine | |
| CN110490220A (en) | A kind of bus load discrimination method and system |
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: 20190129 |