CN109543976A - A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models - Google Patents
A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models Download PDFInfo
- Publication number
- CN109543976A CN109543976A CN201811352980.6A CN201811352980A CN109543976A CN 109543976 A CN109543976 A CN 109543976A CN 201811352980 A CN201811352980 A CN 201811352980A CN 109543976 A CN109543976 A CN 109543976A
- Authority
- CN
- China
- Prior art keywords
- gmm
- wind turbines
- gauss
- data
- harmonic emissions
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 9
- 238000012512 characterization method Methods 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 238000007476 Maximum Likelihood Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 1
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 1
- 235000008434 ginseng Nutrition 0.000 claims 1
- 230000015572 biosynthetic process Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 abstract description 2
- 238000011156 evaluation Methods 0.000 abstract description 2
- 238000003786 synthesis reaction Methods 0.000 abstract description 2
- 238000009826 distribution Methods 0.000 description 5
- 230000005611 electricity Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 239000000203 mixture Substances 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- 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
-
- 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)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- Water Supply & Treatment (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Game Theory and Decision Science (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Probability & Statistics with Applications (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Wind Motors (AREA)
Abstract
The invention discloses a kind of Wind turbines harmonic emissions modal identification methods using gauss hybrid models, comprising: step 1, establishes GMM and initializes the weighted value ω of GMMk, mean μk, covariance matrix ΣkWith subclass number；Step 2, the weighted value ω of GMM is calculated based on EM algorithm iterationk, mean μk, covariance matrix ΣkParameter optimal value, until reaching the local minimum of MDL criterion function；Step 3, according to the parameter optimal value clustering of Wind turbines GMM, class similar in similarity is merged, redundant data is screened out, reduce the subclass number of GMM；Step 4, judge whether subclass number is equal to 1, if it is not, then return step 3, if so, exporting the model result of identification and terminating program.It has the following advantages: the calculation amount for having greatly reduced the scale of data, having reduced comprehensive assessment, but has little influence on the accuracy of electric energy measuring equipment quality synthesis evaluation.
Description
Technical field
The present invention relates to a kind of Wind turbines harmonic emissions modal identification methods using gauss hybrid models.
Background technique
In recent years, the fast development of electric electronic current change technology promotes the continuous increase of new energy development scale, with wind
Power power generation is that the distributed new of representative constantly accesses traditional power grid, also to the humorous of power grid while playing an important role
Wave research brings more challenges.In addition to some traditional problems that harmonic wave causes, since wind power output power is by natural item
The influence of part and there is intermittent and fluctuation, will lead to the fluctuation enhancing of system harmonics, resonance possibility increases, cause wind-powered electricity generation
Unit off-grid.Wind turbines resonance is mainly related with electrical network parameter to unit operational modal, when electric network composition is basically unchanged, wind
The harmonic emissions characteristic (including typical Mode composition, modal parameter and Mode-switch mode) of motor group, is determined whether to occur
The principal element of resonance.The typical Mode of Wind turbines harmonic emissions is thus distinguished from measured data, and then according to modal parameter
Implement optimum control, be the necessary technology means for guaranteeing the consumption of wind-powered electricity generation maximum, to administering Wind turbines harmonic pollution and wind-powered electricity generation simultaneously
Net operation has very important practical significance.
Nowadays, the Power Quality Monitoring Technology of generation of electricity by new energy is highly developed, and a large amount of power qualities of acquisition are gone through
History monitoring data help to analyze the operation characteristic of Wind turbines.Electric energy quality monitoring data have multidimensional, multi-period characteristic, base
Typical Mode is carried out to Wind turbines in such data and excavates the means needed by data mining.Clustering has as one kind
The tool of the data processing of effect can reasonably classify to mass data collection, there is now some Wind turbines harmonic emissions mode
The method of classification, such as K mean cluster etc., main thinking are the basic electric energy quality monitoring data matrixes of construction, are carried out
Raw data matrix coordinate transform finds out the information that several generalized variables characterize former variable, finally uses common clustering algorithm
Carry out the classification of harmonic emission level mode.But the means that generalized variable extracts do not consider initial data characteristic distributions, to extraction
The feature of Wind turbines harmonic emissions model lacks research.
Summary of the invention
The present invention provides a kind of Wind turbines harmonic emissions modal identification methods using gauss hybrid models, overcome
The deficiencies in the prior art described in background technology.
The technical solution adopted by the present invention to solve the technical problems is:
A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models, it includes:
Step 1, it establishes GMM and initializes the weighted value ω of GMMk, mean μk, covariance matrix ΣkWith subclass number；GMM
Expression formula be expressed as follows with the weighting of N number of gauss component:
In formula: ωkIndicate the weight of k-th of multidimensional list Gaussian Profile, pk(x；μk；Σk) indicate k-th of multidimensional list Gauss
The probability density function of distribution；X=(x1,x2,...,xn)TIndicate data sample column vector, which is characterization wind turbine
The characteristic of group harmonic emissions characteristic, wherein each xi(1≤i≤n) indicates that a point in p dimension coordinate system, coordinate are
(α1,α2,…,αp)；N is total sample number strong point to be clustered number, μkIndicate the Mean Vector of k-th of Gauss model, ΣkIndicate the
The variance of k Gauss model；
Step 2, the weighted value ω of GMM is calculated based on EM algorithm iterationk, mean μk, covariance matrix ΣkParameter it is optimal
Value, until reaching the local minimum of MDL criterion function:
Step 3, according to the parameter optimal value clustering of Wind turbines GMM, class similar in similarity is merged, is screened out superfluous
Remainder evidence reduces the subclass number of GMM；
Step 4, judge whether subclass number is equal to 1, if it is not, then return step 3, if so, the model knot of output identification
Fruit simultaneously terminates program.
Among one embodiment: the characteristic of the characterization Wind turbines harmonic emissions characteristic is by Wind turbines harmonic wave
Monitoring system marks the assessment data and national standard limit using the statistical value of continuous measurement period data as assessment data
It is obtained after change processing.
Among one embodiment: the estimating step of the EM algorithm includes:
Step 11, it enables, ι=0,Random selection initializationWith
Define i-th of sample point xiBelong to the initialization posterior probability of kth class Gauss model:
Step 12, result obtained in the previous step is substituted into the GMM cluster that following maximum likelihood formula calculates+1 iteration of ι
Parameter:
Step 13, if | | Φ(l+1)-Φ(l)| | < δ, iteration terminate, Φ(l+1)The GMM parameter as estimated, δ are setting
Iteration ends threshold value；Otherwise, ι=ι+1, and return step 12 are enabled.
The technical program compared with the background art, it has the following advantages:
The present invention proposes that, based on gauss hybrid models (Gaussian Mixture Model, GMM) clustering algorithm, it uses more
The weighted array of a Gaussian Profile probability density function describes multidimensional data vector in the distribution situation of probability space, with tradition
K- mean cluster, the methods of K- central cluster compares, the limitation not being distributed by particular probability, can be on the basis for distinguishing classification
On immediately arrive at the statistical distributions of data, there is good numeracy skills, and can be fitted by increasing model component any
It is continuously distributed.
The present invention proposes a kind of GMM clustering method of Wind turbines harmonic emissions mode.Wind turbines harmonic wave is supervised
After measured data mark change processing, GMM model parameter is calculated with desired maximum value algorithm iteration, it is then poly- to GMM model parameter
Alanysis, GMM clustering algorithm have given up redundancy and unessential information in electric energy measuring equipment quality evaluation original index data,
Sample is clustered based on the minimum length that Minimum description length criterion clustering obtains description Wind turbines harmonic emissions, but is retained
The significant data and characteristic feature of original index data.Therefore, GMM clustering algorithm has greatly reduced the scale of data, has reduced
The calculation amount of comprehensive assessment, but has little influence on the accuracy of electric energy measuring equipment quality synthesis evaluation.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and examples.
Fig. 1 is a kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models described in the present embodiment
Step flow chart.
Specific embodiment
Fig. 1 is please referred to, a kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models, it includes:
Step 1, it establishes GMM and initializes the weighted value ω of GMMk, mean μk, covariance matrix ΣkWith subclass number；GMM
Expression formula be expressed as follows with the weighting of N number of gauss component:
In formula: ωkIndicate the weight of k-th of multidimensional list Gaussian Profile, pk(x；μk；Σk) indicate k-th of multidimensional list Gauss
The probability density function of distribution；X=(x1,x2,...,xn)TIndicate data sample column vector, which is characterization wind turbine
The characteristic of group harmonic emissions characteristic, wherein each xi(1≤i≤n) indicates that a point in p dimension coordinate system, coordinate are
(α1,α2,…,αp)；N is total sample number strong point to be clustered number, μkIndicate the Mean Vector of k-th of Gauss model, ΣkIndicate the
The variance of k Gauss model；
Step 2, the weighted value ω of GMM is calculated based on EM algorithm iterationk, mean μk, covariance matrix ΣkParameter it is optimal
Value, until reaching the local minimum of MDL criterion function:
Step 3, according to the parameter optimal value clustering of Wind turbines GMM, class similar in similarity is merged, is screened out superfluous
Remainder evidence reduces the subclass number of GMM；
Step 4, judge whether subclass number is equal to 1, if it is not, then return step 3, if so, the model knot of output identification
Fruit simultaneously terminates program.
In the present embodiment, the characteristic of the characterization Wind turbines harmonic emissions characteristic is supervised by Wind turbines harmonic wave
Examining system is using the statistical value (generally taking the big value of 95% probability) of continuous measurement period (generally 10min) data as assessment number
According to, and the assessment data and national standard limit marked and is obtained after change is handled.
The estimating step of EM (Expectation Maximum, the EM) algorithm includes:
Step 11, it enables, ι=0,Random selection initializationWith
Define i-th of sample point xiBelong to the initialization posterior probability of kth class Gauss model:
Step 12, result obtained in the previous step is substituted into the GMM cluster that following maximum likelihood formula calculates+1 iteration of ι
Parameter:
Step 13, if | | Φ(l+1)-Φ(l)| | < δ, iteration terminate, Φ(l+1)The GMM parameter as estimated, δ are setting
Iteration ends threshold value；Otherwise, ι=ι+1, and return step 12 are enabled.
When being clustered the present invention is based on GMM, minimum description length (Minimum Description is used
Length, MDL) criterion function construct its cluster objective function.The main thought of MDL is to be built with probabilistic model to object
When mould, the order of accuarcy of model should being considered, while making model most simple again, independent number of parameters is minimum.It can explain
To give a hypothesis set, a data sequence d, trial searching wherein specifically assumes or the combination of certain hypothesis comes
Minimize ground compressed data sequences d.
Cluster Program extracts the mixed model of each data set, regards each model as a small subclass.It is calculated with EM
Method iterates to calculate the weighted value w of each subclassk, mean μkWith covariance ΣkOptimal value, to each signal characteristic data construct
Then each GMM is initialized as independent class by GMM.Constantly iterative solution GMM parameter optimal value, until MDL function is local most
It is small.According to the similarity of GMM optimized parameter, the similar class of similarity is merged, until gathering for a class, two subclasses of every combination
Afterwards, then the class number of former cluster sample is replaced.
The present invention distinguishes the typical Mode of Wind turbines harmonic emissions, Jin Ergen from measured data by GMM clustering
Implement optimum control according to modal parameter, is the necessary technology means for guaranteeing the consumption of wind-powered electricity generation maximum, it is dirty to Wind turbines harmonic wave is administered
Dye and wind-electricity integration operation have very important practical significance.
The above is only the preferred embodiment of the present invention, the range implemented of the present invention that therefore, it cannot be limited according to, i.e., according to
Equivalent changes and modifications made by the invention patent range and description, should still be within the scope of the present invention.
Claims (3)
1. a kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models, it is characterised in that: include:
Step 1, it establishes GMM and initializes the weighted value ω of GMMk, mean μk, covariance matrix ΣkWith subclass number；The table of GMM
It is expressed as follows up to formula with the weighting of N number of gauss component:
In formula: ωkIndicate the weight of k-th of multidimensional list Gaussian Profile, pk(x；μk；Σk) indicate k-th of multidimensional list Gaussian Profile
Probability density function；X=(x1,x2,...,xn)TIndicate data sample column vector, which is characterization Wind turbines harmonic wave
The characteristic of emission characteristics, wherein each xi(1≤i≤n) indicates that a point in p dimension coordinate system, coordinate are (α1,
α2,…,αp)；N is total sample number strong point to be clustered number, μkIndicate the Mean Vector of k-th of Gauss model, ΣkIt indicates k-th
The variance of Gauss model；
Step 2, the weighted value ω of GMM is calculated based on EM algorithm iterationk, mean μk, covariance matrix ΣkParameter optimal value, directly
To the local minimum for reaching MDL criterion function:
Step 3, according to the parameter optimal value clustering of Wind turbines GMM, class similar in similarity is merged, redundant digit is screened out
According to reducing the subclass number of GMM；
Step 4, judge whether subclass number is equal to 1, if it is not, then return step 3, if so, the model result of output identification is simultaneously
Terminate program.
2. a kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models according to claim 1,
It is characterized by: it is described characterization Wind turbines harmonic emissions characteristic characteristic be by Wind turbines Harmonics Monitoring System with
The statistical value of continuous measurement period data is obtained as assessment data, and after the assessment data are marked change processing with national standard limit
?.
3. a kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models according to claim 1,
It is characterized by: the estimating step of the EM algorithm includes:
Step 11, it enables, ι=0,Random selection initializationWithDefinition the
I sample point xiBelong to the initialization posterior probability of kth class Gauss model:
Step 12, result obtained in the previous step is substituted into the GMM cluster ginseng that following maximum likelihood formula calculates+1 iteration of ι
Number:
Step 13, if | | Φ(l+1)-Φ(l)| | < δ, iteration terminate, Φ(l+1)The GMM parameter as estimated, δ are that the iteration of setting is whole
Only threshold value；Otherwise, ι=ι+1, and return step 12 are enabled.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352980.6A CN109543976A (en) | 2018-11-14 | 2018-11-14 | A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811352980.6A CN109543976A (en) | 2018-11-14 | 2018-11-14 | A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109543976A true CN109543976A (en) | 2019-03-29 |
Family
ID=65847369
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811352980.6A Pending CN109543976A (en) | 2018-11-14 | 2018-11-14 | A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109543976A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563623A (en) * | 2020-04-30 | 2020-08-21 | 国网山东省电力公司威海供电公司 | Typical scene extraction method and system for wind power system planning |
-
2018
- 2018-11-14 CN CN201811352980.6A patent/CN109543976A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563623A (en) * | 2020-04-30 | 2020-08-21 | 国网山东省电力公司威海供电公司 | Typical scene extraction method and system for wind power system planning |
| CN111563623B (en) * | 2020-04-30 | 2022-05-10 | 国网山东省电力公司威海供电公司 | Typical scene extraction method and system for wind power system planning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109299812B (en) | Flood prediction method based on deep learning model and KNN real-time correction | |
| AU2016325186A1 (en) | Bus load forecasting method | |
| CN111753893A (en) | Wind turbine generator power cluster prediction method based on clustering and deep learning | |
| CN109670553B (en) | Photovoltaic array fault diagnosis method based on adaptive neural fuzzy inference system | |
| WO2014101579A1 (en) | Method for identifying a transformer local discharge mode based on a singular value decomposition algorithm | |
| CN106485089B (en) | The interval parameter acquisition methods of harmonic wave user's typical condition | |
| CN108428017A (en) | Wind power interval prediction method based on core extreme learning machine quantile estimate | |
| CN109165819B (en) | Active power distribution network reliability rapid evaluation method based on improved AdaBoost. M1-SVM | |
| CN107038292A (en) | A kind of many output of wind electric field correlation modeling methods based on adaptive multivariable nonparametric probability | |
| CN110909977A (en) | Power grid fault diagnosis method based on ADASYN-DHSD-ET | |
| CN108428015A (en) | A kind of wind power forecasting method based on history meteorological data and stochastic simulation | |
| CN112186761B (en) | Wind power scene generation method and system based on probability distribution | |
| CN109086527A (en) | A kind of practical equivalent modeling method based on running of wind generating set state | |
| CN108717579A (en) | A kind of short-term wind-electricity power interval prediction method | |
| CN110738232A (en) | grid voltage out-of-limit cause diagnosis method based on data mining technology | |
| Peng et al. | A very short term wind power prediction approach based on Multilayer Restricted Boltzmann Machine | |
| CN110738253A (en) | short-term wind power prediction method based on FCM and AFSA-Elman | |
| Syu et al. | Ultra-short-term wind speed forecasting for wind power based on gated recurrent unit | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN113222067B (en) | Island intelligent detection method based on SVM-Adaboost algorithm | |
| CN110991689A (en) | Distributed photovoltaic power generation system short-term prediction method based on LSTM-Morlet model | |
| CN109858667A (en) | It is a kind of based on thunder and lightning weather to the short term clustering method of loading effects | |
| CN109543976A (en) | A kind of Wind turbines harmonic emissions modal identification method using gauss hybrid models | |
| Lingqing et al. | Detection method for power theft based on SOM neural network and K-means clustering algorithm | |
| CN115659258B (en) | Power distribution network fault detection method based on multi-scale graph roll-up twin network |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190329 |