CN107491412A - A kind of user power utilization load characteristic extracting method based on experience wavelet transformation - Google Patents
A kind of user power utilization load characteristic extracting method based on experience wavelet transformation Download PDFInfo
- Publication number
- CN107491412A CN107491412A CN201710557184.5A CN201710557184A CN107491412A CN 107491412 A CN107491412 A CN 107491412A CN 201710557184 A CN201710557184 A CN 201710557184A CN 107491412 A CN107491412 A CN 107491412A
- Authority
- CN
- China
- Prior art keywords
- experience
- user
- power utilization
- experience wavelet
- user power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a kind of user power utilization load characteristic extracting method based on experience wavelet transformation, including:It is multiple experience wavelet function components to decompose user power utilization load data with experience wavelet transformation;Hilbert conversion is carried out to each experience wavelet function, and Hilbert spectrums are indicated in time-frequency combination domain;Calculate each component and original loads data normalization coefficient correlation, suitable threshold value is set according to degree of correlation selection rule, choose and the higher active constituent of original loads Data mutuality degree, the energy of active constituent to being extracted carries out 2 norm solutions, and solving result is combined to the characteristic vector for forming energy comprehensive characterization user power utilization characteristic.This method combines the advantages of wavelet analysis and ensemble empirical mode decomposition method, by adaptively splitting Fourier spectrum, orthogonal wavelet filter group is constructed, single signal is decomposed into multiple amplitude modulationfrequency modulation components with compact schemes fourier spectrum, realizes the separation of different frequency characteristic information.
Description
Technical field
The present invention relates to power system customer power load feature extraction field, more particularly to a kind of experience small echo that is based on to become
The user power utilization load characteristic extracting method changed.
Background technology
Increasing sensor, intelligence instrument are loaded into intelligent distribution network at present, for obtaining the reality of electric power networks
When data.Wherein, user power utilization load data is played an important role in terms of the formulation of operation of power networks decision-making, therefore electricity consumption is born
Lotus data carry out Study on Feature Extraction for supporting the safety, economy, reliability service of power network to have important theory significance and reality
Border is worth.
The conventional method of signal characteristic abstraction has at present:Fourier transformation, wavelet analysis method, analysis method of wavelet packet and
Hilbert-Huang transform etc..Fourier transformation uses global sinusoidal quantity, though it is different signal effectively can be divided into frequency
Each component, but the missing of locality information can be caused.Although wavelet and wavelet packets conversion can be used for analyzing non-stationary signal,
The select permeability for solving wavelet basis is also needed during practical application, does not possess adaptivity.It is related in Hilbert-Huang transform method
Empirical mode decomposition algorithm, before compensate for method dependent on priori choose basic function the drawbacks of, but this algorithm lack have
The mathematical theory of power supports and mode mixing phenomenon occurs in signal processing.Experience small wave converting method solves
The problem of theoretical foundation deficiency and algorithm adaptability, it is intended to build one by the directly caused basic function of characteristics of signals, Jin Ertong
One appropriate wavelet filter group pattern different to signal of design is crossed to extract.
The content of the invention
It is an object of the present invention to give a kind of power load feature extracting method based on experience wavelet transformation, it is
Analysis user power utilization characteristic provides effective means.
The present invention relates to a kind of user power utilization load characteristic extracting method based on experience wavelet transformation, including it is as follows
Step:
Decomposed, obtained more using experience small wave converting method using the original power load data of user as primary signal
Individual experience wavelet function component;
By carrying out Hilbert conversion to each experience wavelet function component of the acquisition, obtain characterizing each component
Instantaneous amplitude and the Hilbert of instantaneous frequency spectrum, and by its Hilbert spectrum when-frequency combine domain in be indicated, so as to establish
The time-frequency representation of the original power load data-signal of user;
Calculate the normalizated correlation coefficient of the original power load data of user and each experience wavelet function component, filter out with
The higher active constituent of the original power load Data mutuality degree of user;
User power utilization load data characteristic vector is built according to active constituent, this characteristic vector is original as identification user
Power load feature exports.
Further, Fourier transformation is carried out to primary signal and obtains its Fourier frequency spectrum, by Fourier frequency spectrums
Adaptivenon-uniform sampling constructs experience wavelet filter group, AM/FM amplitude modulation/frequency modulation (AM-FM) signal of extraction with compact schemes fourier spectrum
Composition, the component for including different frequency characteristic information in signal component is separated, obtain multiple covering different frequency feature
The experience wavelet function component of information.
Wherein, screening is to set suitable threshold value according to degree of correlation selection rule to choose degree of correlation height, can reflect use
The active constituent of family power load feature, reach the purpose for removing pseudo- component and unrelated interruptions item.
Further, the method that power load data characteristics vector is built according to active constituent is each active constituent of calculating
Energy, and 2 norm solutions are carried out to the energy, solving result is combined, constructing can comprehensive characterization user power utilization spy
The characteristic vector of property.
In addition, when normalizated correlation coefficient is more than or equal to 0.5, it is considered as degree of correlation height.
Brief description of the drawings
Fig. 1 is a kind of step frame of user power utilization load characteristic extracting method based on experience wavelet transformation of the present invention
Figure;
Fig. 2 is the original user power load schematic diagram data of sensor collection;
Fig. 3 is that original user power load data-signal is carried out using experience wavelet transformation to pass through caused by adaptive decomposition
Test wavelet function component schematic diagram;
Fig. 4 is time-frequency representation schematic diagram of each experience wavelet function component after Hilbert is converted.
Embodiment
The specific implementation step of the present invention is described further with reference to real case and accompanying drawing.
A kind of user power utilization load characteristic extracting method based on experience wavelet transformation:
Step 1, user power utilization load data is decomposed using experience wavelet transformation.
S1:Fig. 2 is the original user power load data of sensor collection, with a time domain discrete series X=[x1,x2,…,
xN] be indicated.Fourier transformation is carried out to this sequence, obtains Fourier spectrum F (ω).
S2:Fourier is composed and carries out adaptivenon-uniform sampling to construct suitable wavelet filter group.Frequency spectrum division principle is such as
Under:
Assuming that with ωnFrequency spectrum F (ω) is divided into N number of successive passage on [0, π] for border, then frequency spectrum division includes N+
1 border, remove border ω0=0, ωN=π, the determination method on remaining N-1 border are:The local pole in frequency spectrum is detected first
Big value simultaneously sorts it (including 0 and π) in descending order.Assuming that M maximum be present:
If M >=N, illustrate to have enough maximum to define spectral regions segmentation, then N-1 maximum before retaining;
If M < N, illustrate that the decomposition mode of actual data signal is less than desired value, whole maximum need to be retained and reset N
To suitable value.
The intermediate frequency of two continuous maximum is taken as ωn, with ω0=0, ωN=π collectively forms final frequency spectrum and drawn
Divide result.
S3:It is segmented according to Fourier spectrum in S2, constructs N number of experience small echo, determines experience wavelet function component.Establish warp
Test wavelet basis, experience scaling functionWith experience wavelet mother functionIt is defined as follows:
In formula:Function β (x) is an arbitrary Ck([0,1]) function is, it is necessary to meet following condition:
And
τn=γ ωn
Experience wavelet transformation is obtained by the inner product of experience small echo:
Approximation coefficient is obtained by the inner product with zoom function:
Signal f (t) experience wavelet transformation construction is as follows:
The pattern f that experience is decomposedk, it is given by:
Original user power load data-signal is decomposed into by the wavelet filter group according to frequency spectrum adaptivenon-uniform sampling construction
Multiple experience wavelet function components for covering different frequency characteristic information, as shown in Figure 3.
Step 2, Hilbert conversion is carried out respectively to N+1 experience wavelet function, establishes the time-frequency representation of each component.
Hilbert conversion is carried out to each experience wavelet function component, conversion is as follows:
Wherein (p. υ) is Cauchy's principal value, and f (t) is experience wavelet function component.
Then the analytical form of signal is as follows:
fa(t)=f (t)+ι Hf(t)
Wherein AM-FM signal forms are:
Then signal resolution form is represented by:
Thus the instantaneous amplitude F (t) and instantaneous frequency of each experience wavelet function be can extract outFig. 4 gives each warp
Test time-frequency representation of the wavelet function after Hilbert is converted.
Step 3, extraction and the higher active constituent of the original loads data-signal degree of correlation.
Calculate the normalizated correlation coefficient r between each experience wavelet function and power load data-signalj, rj(j=1,
2 ..., n) calculation formula be:
Wherein xiFor i-th of element of power load signal sequence, x is power load signal sequence average, fiFor j-th
I-th of element of experience wavelet function component,For j-th of experience wavelet function component average.
Under normal circumstances, when normalizated correlation coefficient is more than or equal to 0.5, it is considered as height correlation, according to this degree of correlation
Selection rule, choose normalizated correlation coefficient rj>=0.5 component solves as active constituent for characteristic vector.It is computed,
4 experience wavelet functions are chosen as active constituent.
Step 4, construction feature vector.
S1:The ENERGY E of each active constituent extracted in calculation procedure 3j。
Ej=∫ | fj(t)|2dt
S2:2 norm calculations are carried out to each energy.
S3:By the characteristic vector V of result of calculation composite construction energy comprehensive characterization user power utilization characteristic.
V=[v1,v2,...,vN]
It is computed, the characteristic vector of the user power utilization load data is V=[21.29 19.67 18.57 11.89].
Claims (5)
1. a kind of user power utilization load characteristic extracting method based on experience wavelet transformation, including step:
Decomposed using the original power load data of user as primary signal using experience small wave converting method, obtain multiple warps
Test wavelet function component;
By carrying out Hilbert conversion to each experience wavelet function component of the acquisition, obtain characterizing each component instantaneous
The Hilbert of amplitude and instantaneous frequency compose, and by its Hilbert spectrum when-frequency combine domain in be indicated, so as to establish user
The time-frequency representation of original power load data-signal;
The normalizated correlation coefficient of the original power load data of user and each experience wavelet function component is calculated, is filtered out and user
The higher active constituent of original power load Data mutuality degree;
User power utilization load data characteristic vector is built according to active constituent, using this characteristic vector as the identification original electricity consumption of user
Load characteristic exports.
2. a kind of user power utilization load characteristic extracting method based on experience wavelet transformation according to claim 1, it is special
Sign is:The specific method for obtaining multiple experience wavelet function components is:Fourier transformation is carried out to the primary signal to obtain
Its Fourier frequency spectrum, by constructing experience wavelet filter group to the adaptivenon-uniform sampling of Fourier frequency spectrums, extraction has tight branch
AM/FM amplitude modulation/frequency modulation (AM-FM) signal component of fourier spectrum is supportted, the component of different frequency characteristic information will be included in signal component
Separated, obtain multiple experience wavelet function components for covering different frequency characteristic information.
3. a kind of user power utilization load characteristic extracting method based on experience wavelet transformation according to claim 2, it is special
Sign is:The screening is to set suitable threshold value according to degree of correlation selection rule to choose degree of correlation height, can reflect user
The active constituent of power load feature, reach the purpose for removing pseudo- component and unrelated interruptions item.
4. a kind of user power utilization load characteristic extracting method based on experience wavelet transformation according to claim 3, it is special
Sign is that the method that power load data characteristics vector is built according to active constituent is the energy of each active constituent of calculating,
And 2 norm solutions are carried out to the energy, solving result is combined, constructs the spy of energy comprehensive characterization user power utilization characteristic
Levy vector.
5. a kind of user power utilization load characteristic extracting method based on experience wavelet transformation according to claim 3, it is special
Sign is, when normalizated correlation coefficient is more than or equal to 0.5, to be considered as degree of correlation height.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710557184.5A CN107491412B (en) | 2017-07-10 | 2017-07-10 | User electricity load characteristic extraction method based on empirical wavelet transform |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710557184.5A CN107491412B (en) | 2017-07-10 | 2017-07-10 | User electricity load characteristic extraction method based on empirical wavelet transform |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107491412A true CN107491412A (en) | 2017-12-19 |
CN107491412B CN107491412B (en) | 2021-01-12 |
Family
ID=60644352
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710557184.5A Expired - Fee Related CN107491412B (en) | 2017-07-10 | 2017-07-10 | User electricity load characteristic extraction method based on empirical wavelet transform |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107491412B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109269629A (en) * | 2018-08-03 | 2019-01-25 | 河海大学 | Paralleling reactor of extra-high voltage analysis of vibration signal method based on experience wavelet transformation |
CN110490369A (en) * | 2019-07-25 | 2019-11-22 | 国网安徽省电力有限公司 | A kind of Short-Term Load Forecasting Method based on EWT and LSSVM model |
CN112200384A (en) * | 2020-10-28 | 2021-01-08 | 宁波立新科技股份有限公司 | Power load short-time prediction method based on EWT neural network |
CN114937011A (en) * | 2022-05-12 | 2022-08-23 | 北京航空航天大学 | Photovoltaic cell image anomaly detection method based on empirical Garbor wavelet transform |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030171899A1 (en) * | 2003-04-03 | 2003-09-11 | Castellane Raymond M. | Detecting, classifying and localizing minor amounts of an element within a sample of material |
CN102855385A (en) * | 2012-07-31 | 2013-01-02 | 上海交通大学 | Wind power generation short-period load forecasting method |
CN103390117A (en) * | 2013-08-08 | 2013-11-13 | 山东大学 | Feature extracting method for power load dynamic features |
CN103822786A (en) * | 2012-11-16 | 2014-05-28 | 中国水利电力物资有限公司 | Wind turbine generator set mechanical equipment state diagnosis method based on multivariate statistical analysis |
-
2017
- 2017-07-10 CN CN201710557184.5A patent/CN107491412B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030171899A1 (en) * | 2003-04-03 | 2003-09-11 | Castellane Raymond M. | Detecting, classifying and localizing minor amounts of an element within a sample of material |
CN102855385A (en) * | 2012-07-31 | 2013-01-02 | 上海交通大学 | Wind power generation short-period load forecasting method |
CN103822786A (en) * | 2012-11-16 | 2014-05-28 | 中国水利电力物资有限公司 | Wind turbine generator set mechanical equipment state diagnosis method based on multivariate statistical analysis |
CN103390117A (en) * | 2013-08-08 | 2013-11-13 | 山东大学 | Feature extracting method for power load dynamic features |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109269629A (en) * | 2018-08-03 | 2019-01-25 | 河海大学 | Paralleling reactor of extra-high voltage analysis of vibration signal method based on experience wavelet transformation |
CN110490369A (en) * | 2019-07-25 | 2019-11-22 | 国网安徽省电力有限公司 | A kind of Short-Term Load Forecasting Method based on EWT and LSSVM model |
CN112200384A (en) * | 2020-10-28 | 2021-01-08 | 宁波立新科技股份有限公司 | Power load short-time prediction method based on EWT neural network |
CN112200384B (en) * | 2020-10-28 | 2024-05-17 | 宁波立新科技股份有限公司 | EWT neural network-based short-time prediction method for power load |
CN114937011A (en) * | 2022-05-12 | 2022-08-23 | 北京航空航天大学 | Photovoltaic cell image anomaly detection method based on empirical Garbor wavelet transform |
CN114937011B (en) * | 2022-05-12 | 2024-05-28 | 北京航空航天大学 | Photovoltaic cell image anomaly detection method based on empirical Garbor wavelet transformation |
Also Published As
Publication number | Publication date |
---|---|
CN107491412B (en) | 2021-01-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107491412A (en) | A kind of user power utilization load characteristic extracting method based on experience wavelet transformation | |
Nguyen et al. | Power quality disturbance classification utilizing S-transform and binary feature matrix method | |
CN109975673B (en) | Method for identifying fault arc at direct current side of photovoltaic microgrid | |
Reddy et al. | A DOST based approach for the condition monitoring of 11 kV distribution line insulators | |
CN105004939B (en) | A kind of complex electric energy quality disturbance signal quantitative analysis method | |
Gai et al. | Multiscale texture classification using reduced quaternion wavelet transform | |
CN102176779B (en) | Wireless multimedia sensing network video signal adaptive sampling and spectrum allocation method | |
CN110119493A (en) | A kind of Power System Events cognitive method based on compression PMU data and the local factor that peels off | |
CN109886908A (en) | Infrared image and visible light image fusion method | |
CN107315111A (en) | A kind of Power Quality Disturbance Classification Method and system | |
CN112819649A (en) | Method and device for determining station area subscriber change relationship | |
CN108108659B (en) | Island detection key feature extraction method based on empirical mode decomposition | |
CN113723244A (en) | Radar radiation source signal separation method based on improved variational modal decomposition | |
Upadhyaya et al. | Evaluation of power quality disturbance in PV-connected IEEE-14 bus system using lifting-based wavelet transform and random forest | |
Krishna et al. | Image pattern recognition technique for the classification of multiple power quality disturbances | |
Liu et al. | Multimedia image compression method based on biorthogonal wavelet and edge intelligent analysis | |
Li et al. | Instructions data compression for smart grid monitoring using wavelet domain singular value decomposition | |
CN115293303A (en) | High-voltage transmission line network monitoring method, system, equipment and medium | |
Sima et al. | A framework for automatically cleansing overvoltage data measured from transmission and distribution systems | |
Beg et al. | A novel Clarke wavelet transform method to classify power system disturbances | |
Gök et al. | Transient event classification using pmu data with deep learning techniques and synthetically supported training‐set | |
CN114139586A (en) | Electric energy quality composite disturbance identification method based on multi-scale analysis and two-dimensional feature extraction | |
Tang et al. | EgeFusion: Towards Edge Gradient Enhancement in Infrared and Visible Image Fusion With Multi-Scale Transform | |
CN106384308A (en) | Power consumption load data processing method and apparatus for power system | |
Li et al. | Pixel level image fusion based the wavelet transform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210112 Termination date: 20210710 |
|
CF01 | Termination of patent right due to non-payment of annual fee |