CN109615142A - A kind of wind farm wind velocity combination forecasting method based on wavelet analysis - Google Patents
A kind of wind farm wind velocity combination forecasting method based on wavelet analysis Download PDFInfo
- Publication number
- CN109615142A CN109615142A CN201811546166.8A CN201811546166A CN109615142A CN 109615142 A CN109615142 A CN 109615142A CN 201811546166 A CN201811546166 A CN 201811546166A CN 109615142 A CN109615142 A CN 109615142A
- Authority
- CN
- China
- Prior art keywords
- wind
- signal
- wind speed
- wavelet analysis
- value
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 34
- 238000013277 forecasting method Methods 0.000 title claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 24
- 239000002245 particle Substances 0.000 claims abstract description 23
- 238000012706 support-vector machine Methods 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims abstract description 18
- 230000002068 genetic effect Effects 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims description 14
- 238000005457 optimization Methods 0.000 claims description 11
- 238000000354 decomposition reaction Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 9
- 230000035772 mutation Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 108090000623 proteins and genes Proteins 0.000 claims description 6
- 241000208340 Araliaceae Species 0.000 claims description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 4
- 235000008434 ginseng Nutrition 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 22
- 230000006870 function Effects 0.000 description 20
- 238000012913 prioritisation Methods 0.000 description 7
- 230000005611 electricity Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013439 planning Methods 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
- G06Q10/04—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
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A30/00—Adapting or protecting infrastructure or their operation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of wind farm wind velocity combination forecasting method based on wavelet analysis, using the combination of the methods of wavelet analysis, genetic algorithm, particle swarm algorithm, neural network and support vector machines, fully consider nonlinearity, randomness and the uneven stability of wind velocity signal, wind velocity signal is decomposed into multilayer signal using wavelet analysis, and prediction is trained to each layer signal by a variety of prediction techniques, the precision of prediction of global convergence precision and wind velocity signal is improved by the configuration of weight coefficient；This method can get compared with the higher forecasting wind speed result of Individual forecast method precision.Improve wind power prediction accuracy for electric system and technical reference is provided, be conducive to dispatching of power netwoks department reasonable arrangement operation plan, reduce operation of power networks cost, guarantees the operation of power grid complete stability.
Description
Technical field
The present invention relates to technical field of new energy power generation, and in particular to a kind of wind farm wind velocity combination based on wavelet analysis
Prediction technique.
Background technique
In recent years, with the fast development of wind energy power technology and utilization, installed capacity of wind-driven power is steeply risen, global wind-powered electricity generation
Industry flourishes.However, wind-power electricity generation is combined with traditional power grid system due to the intermittence and randomness of wind-power electricity generation
Face lot of challenges, including energy power generation planning and turbine service scheduling, the change of network system safe operation and interconnection standard
Change etc..Accurate wind power prediction can provide important evidence for power scheduling, effectively mitigate influence of the wind-powered electricity generation to power grid.Due to
Wind power and wind speed have the relationship directly determined, and wind power prediction can be realized on the basis of forecasting wind speed, so in order to
Mitigate the above problem caused by wind energy access electric system, Accurate Prediction is carried out to short-term wind speed and is become more and more important.
Currently, mainly being predicted in the prior art using single model short-term wind speed, such as Method of Physical Modeling NWP
(numerical weather forecast), statistical learning method (time series method, Kalman filtering method, grey method etc.) and intelligence machine
Learning method (artificial neural network method) etc., but using single model short-term wind speed is predicted when, prediction result be easy by
To the influence of wind speed nonlinearity, and prediction result is easily trapped into local optimum, and checking precision is low, and generalization ability is insufficient,
Precision of prediction is reduced to a certain extent.
Summary of the invention
To solve deficiency in the prior art, the present invention provides a kind of wind farm wind velocity combined prediction based on wavelet analysis
Method solves the essence of the forecasting wind speed due to caused by the randomness of wind farm wind velocity signal and uneven stability etc. in the prior art
Spend not high problem.
In order to achieve the above objectives, the present invention adopts the following technical scheme:
A kind of wind farm wind velocity combination forecasting method based on wavelet analysis, it is characterised in that: operate according to the following steps:
Step 1: carrying out test acquisition to the air speed data in wind field, does normalization operation, history of forming to data sample
Wind series data；Divide the training set and test set of historical wind speed signal；
Step 2: using wavelet analysis method by historical wind speed signal decomposition at n-layer different frequency, the signal of different levels
Component, n are positive integers, indicate the number of plies of signal component；
Step 3: using the initial value and threshold value of Genetic Algorithm Optimized Neural Network, establishing neural network prediction model, right
The low-frequency approximation signal of historical wind speed signal is trained and predicts；The kernel functional parameter g of Support Vector Machines Optimized and punishment because
Sub- C, the support vector machines forecasting wind speed model chosen optimal kernel functional parameter g and penalty factor, establish particle group optimizing,
The high frequency component signal of historical wind speed signal is trained and is predicted；
Step 4: comparing the predicted value and original value of air speed data, and the prediction error of calculation of wind speed signal judges that wind speed is pre-
Whether the error of measured value and true value meets the requirements, and next step is entered to if meeting the requirements, if being unsatisfactory for requiring, by wind
Fast data Decomposition order and mode input dimension optimize as Optimal Parameters further progress；And it is missed according to each prediction signal
Difference determines corresponding weight coefficient；
Step 5: it according to the weight coefficient of multiple forecasting wind speed, reconstructs the windy fast data linear superposition to obtain wind speed pre-
Measured value.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned is special
Sign is: the initial value and threshold value of Genetic Algorithm Optimized Neural Network are utilized in step 3；Specific step is as follows:
Step 1: random initializtion population: individual UVR exposure uses real coding mode, and each individual is a real number
String, the real number string include input layer and hidden layer connection weight, hidden layer threshold value, hidden layer and output layer connection weight and defeated
Layer threshold value out；
Step 2: determine fitness function: with neural network prediction output desired output between absolute error and
Inverse as fitness function F；
Step 3: selection operation: selecting several body as parents for raising up seed from population, high of fitness
Body is genetic to that follow-on probability is larger, and the low individual inheritance of fitness is then smaller to follow-on probability；
Step 4: crossover operation: the individual of two pairings is with crossover probability pcPart of gene is exchanged, forms two newly
Individual；
Step 5: mutation operation: with a smaller mutation probability pvJ-th of gene g of i-th of individual of selectionijInto
Row variation；
Step 6: fitness function value is calculated, the weight and threshold value of optimization are exported if meeting algorithm termination condition, if not
Meet algorithm termination condition then return step three；
Step 7: using the weight of the optimization of genetic algorithm output and threshold value as the initial weight and threshold value of neural network,
With training sample to neural metwork training, short-term wind speed forecasting model is obtained.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned is special
Sign is: in step 3 using particle swarm algorithm Support Vector Machines Optimized wind speed combination forecasting in penalty factor and
Optimal penalty factor and kernel functional parameter g are assigned to support vector machines by kernel functional parameter g, and with the wind speed sample pair of building
Support vector machines network training establishes the wind speed combination forecasting of optimization, analysis and evaluation and foreca result.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned is special
Sign is:
Using the penalty factor and kernel function ginseng in the wind speed combination forecasting of particle swarm algorithm Support Vector Machines Optimized
The specific practice of number g is:
A group particle is initialized in solution space, each particle I represents the penalty factor and kernel function ginseng of built-up pattern
A potential optimal solution of number g, with position vector XI=(xI1,xI2,…xIn) velocity vector VI=(vI1,vI2,…vIn) two n
Dimensional vector indicates, speed and location updating equation are as follows:
In formula,Indicate the particle position and speed that d is tieed up in k iteration Chinese style；W indicates weight；c1, c2
Indicate that Studying factors, Chang Jun take 2,Take the random number between [0,1]；Indicate particle I individual pole
It is worth the coordinate tieed up in d；It is suitable to indicate that entire group's global extreme point searches after successive ignition in the coordinate that d is tieed up
Answer angle value optimal location, as optimal penalty factor and kernel functional parameter g.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned,
It is characterized by: the number of plies n of signal component is equal to 5 in step 2.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned,
It is characterized by: five layer signal components include one layer of low frequency signal and four layers of high-frequency signal.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned is special
Sign right in: the kernel function that step 3 Kernel Function parameter g is used is Gaussian radial basis function k (xi, xj)=exp [- (| |
xi-xj||2)/σ2]；X in formulaiIt is spatially any point, xjIt is kernel function center, σ is the width parameter of function, control function
Radial effect range.
As a kind of prioritization scheme, a kind of wind farm wind velocity combination forecasting method based on wavelet analysis above-mentioned,
It is characterized by: also being carried out to the exceptional value of acquired air speed data flat before the data sample in step 1 does normalization operation
Steadyization processing, studies for a second time courses one has flunked missing data, chooses data sample.
Advantageous effects of the invention: the present invention fully considers the nonlinearity of wind velocity signal, randomness and not
Wind velocity signal is decomposed into multilayer signal using wavelet analysis, and is carried out by a variety of prediction techniques to each layer signal by stationarity
Training prediction, the precision of prediction of global convergence precision and wind velocity signal is improved by the configuration of weight coefficient；This method can obtain
It obtains compared with the higher forecasting wind speed result of Individual forecast method precision.Improve wind power prediction accuracy for electric system and skill is provided
Reference in art is conducive to dispatching of power netwoks department reasonable arrangement operation plan, reduces operation of power networks cost, guarantees that power grid is completely steady
Fixed operation.
Detailed description of the invention
Fig. 1 is the forecasting wind speed flow chart of present invention combination wind speed forecasting method；
Fig. 2 is historical wind speed data used by the embodiment of the present invention；
Fig. 3 is the wind velocity signal after wavelet decomposition；
Fig. 4 is the wind speed value of the embodiment of the present invention and the comparison diagram of test value.
Specific embodiment
The invention will be further described below in conjunction with the accompanying drawings.Following embodiment is only used for clearly illustrating the present invention
Technical solution, and not intended to limit the protection scope of the present invention.
Present embodiment discloses a kind of wind farm wind velocity combination forecasting method based on wavelet analysis, including wavelet analysis,
The combination of the methods of genetic algorithm, particle swarm algorithm, neural network and support vector machines.It is effective to verify the technology of the present invention method
The accuracy of property and forecasting wind speed, establishes forecasting wind speed model with the wind speed test data in somewhere, and be trained and predict.Figure
1 is a kind of flow diagram of the wind farm wind velocity combination forecasting method based on wavelet analysis provided in an embodiment of the present invention.Its
Specific implementation the following steps are included:
Step 1: carrying out test acquisition to the air speed data in wind field first, does normalization operation to data sample, is formed
Historical wind speed sequence data；Divide the training set and test set of historical wind speed signal.For promoted test result accuracy, in number
Before doing normalization operation according to sample, tranquilization processing preferably is carried out to the exceptional value of acquired air speed data, studies for a second time courses one has flunked missing data,
Data sample is chosen, the historical wind speed after handling is as shown in Figure 2.
Step 2: then using wavelet analysis method by historical wind speed signal decomposition at n-layer different frequency, different levels
Signal component, n are positive integers, indicate the number of plies of signal component.The number of plies n of the signal component of the present embodiment is preferably equal to 5, including
One layer of low frequency signal and four layers of high-frequency signal, the exploded relationship of concrete signal are as follows: S=a4+d4+d3+d2+d1；A indicates low frequency
Signal, d indicate that high-frequency signal, S indicate historical wind speed signal.Training set data is reconstructed, forecasting wind speed training mould is formed
Type is output and input；Wind velocity signal after wavelet decomposition is as shown in Figure 3.
Step 3: the initial value and threshold value of optimization neural network establish neural network prediction model, to historical wind speed signal
Low-frequency approximation signal be trained and predict；The kernel functional parameter g and penalty factor of Support Vector Machines Optimized choose optimal
Kernel functional parameter g and penalty factor, the support vector machines forecasting wind speed model for establishing particle group optimizing, to historical wind speed believe
Number high frequency component signal be trained and predict, the initial value and threshold using Genetic Algorithm Optimized Neural Network in this step
Value；Specific step is as follows: random initializtion population: individual UVR exposure uses real coding mode, and each individual is a real number
String, the real number string include input layer and hidden layer connection weight, hidden layer threshold value, hidden layer and output layer connection weight and defeated
Layer threshold value out；Determine fitness function: with the inverse of the absolute error sum between the prediction output of neural network and desired output
As fitness function F；Selection operation: select several body as parents for raising up seed from population, fitness is high
Individual inheritance is larger to follow-on probability, and the low individual inheritance of fitness is then smaller to follow-on probability；Crossover operation: two
The individual of a pairing is with crossover probability pcPart of gene is exchanged, two new individuals are formed；Mutation operation: with a comparison
Small mutation probability pvJ-th of gene g of i-th of individual of selectionijIt makes a variation；Fitness function value is calculated, if meeting algorithm
Termination condition then exports the weight and threshold value of optimization, and the selection behaviour in this step is returned to if discontented afc algorithm termination condition
Make link；Using the weight of the optimization of genetic algorithm output and threshold value as the initial weight and threshold value of neural network, with training sample
This obtains short-term wind speed forecasting model to neural metwork training.
In this step using particle swarm algorithm Support Vector Machines Optimized wind speed combination forecasting in penalty factor and
Optimal penalty factor and kernel functional parameter g are assigned to support vector machines by kernel functional parameter g, and with the wind speed sample pair of building
Support vector machines network training establishes the wind speed combination forecasting of optimization, analysis and evaluation and foreca result.It is preferred that using particle
The specific practice of penalty factor and kernel functional parameter g in the wind speed combination forecasting of group's algorithm optimization support vector machines is:
A group particle is initialized in solution space, each particle I represents the penalty factor and kernel function ginseng of built-up pattern
A potential optimal solution of number g, with position vector XI=(xI1,xI2,…xIn) velocity vector VI=(vI1,vI2,…vIn) two n
Dimensional vector indicates, speed and location updating equation are as follows:
In formula,Indicate the particle position and speed that d is tieed up in k iteration Chinese style；W indicates weight；c1, c2
Indicate that Studying factors, Chang Jun take 2,Take the random number between [0,1]；Indicate particle I individual pole
It is worth the coordinate tieed up in d；It is suitable to indicate that entire group's global extreme point searches after successive ignition in the coordinate that d is tieed up
Answer angle value optimal location, as optimal penalty factor and kernel functional parameter g.
It is preferred that: the kernel function that kernel functional parameter g is used is Gaussian radial basis function K (xi, xj)=exp [- (| | x-xi|
|2)/σ2]；X in formulaiIt is spatially any point, xjIt is kernel function center, σ is the width parameter of function, and the radial of control function makees
Use range.
Step 4: comparing the predicted value and original value of air speed data, and the prediction error of calculation of wind speed signal judges that wind speed is pre-
Whether the error of measured value and true value meets the requirements, and next step is entered to if meeting the requirements, if being unsatisfactory for requiring, by wind
Fast data Decomposition order and mode input dimension optimize as Optimal Parameters further progress；And it is missed according to each prediction signal
Difference determines corresponding weight coefficient；The bigger condition of the smaller equivalent layer weight coefficient of error need to be met, such as: 1 error of signals layer is
e1, 2 error of signals layer is e2, then the weight coefficient of corresponding each layer is respectively λ1=e2/(e1+e2)；λ2=e1/(e1+e2)。
Step 5: it according to the weight coefficient of multiple forecasting wind speed, reconstructs the windy fast data linear superposition to obtain wind speed pre-
Measured value.As shown in figure 4, it can be found that using the obtained wind speed value of the method for the present invention and true value error very little, effectively
Verify the validity and accuracy of forecast of the method for the present invention.
The present invention fully considers nonlinearity, randomness and the uneven stability of wind velocity signal, using wavelet analysis by wind
Fast signal decomposition is multilayer signal, and is trained prediction to each layer signal by a variety of prediction techniques, passes through weight coefficient
Configuration improves the precision of prediction of global convergence precision and wind velocity signal；This method can get higher compared with Individual forecast method precision
Forecasting wind speed result.Improve wind power prediction accuracy for electric system and technical reference is provided, is conducive to power grid tune
Degree department reasonable arrangement operation plan reduces operation of power networks cost, guarantees the operation of power grid complete stability.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the technical principles of the invention, several improvement and deformations can also be made, these improvement and deformations
Also it should be regarded as protection scope of the present invention.
Claims (8)
1. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis, it is characterised in that: operate according to the following steps:
Step 1: carrying out test acquisition to the air speed data in wind field, does normalization operation, history of forming wind speed to data sample
Sequence data；Divide the training set and test set of historical wind speed signal；
Step 2: using wavelet analysis method by historical wind speed signal decomposition at the signal point of n-layer different frequency, different levels
Amount, n is positive integer, indicates the number of plies of signal component；
Step 3: using the initial value and threshold value of Genetic Algorithm Optimized Neural Network, neural network prediction model is established, to history
The low-frequency approximation signal of wind velocity signal is trained and predicts；The kernel functional parameter g and penalty factor of Support Vector Machines Optimized,
The support vector machines forecasting wind speed model chosen optimal kernel functional parameter g and penalty factor, establish particle group optimizing, to going through
The high frequency component signal of history wind velocity signal is trained and predicts；
Step 4: comparing the predicted value and original value of air speed data, and the prediction error of calculation of wind speed signal judges wind speed value
Whether met the requirements with the error of true value, next step is entered to if meeting the requirements, if being unsatisfactory for requiring, by wind speed number
Optimize according to Decomposition order and mode input dimension as Optimal Parameters further progress；And it is true according to each prediction signal error
Fixed corresponding weight coefficient；
Step 5: it according to the weight coefficient of multiple forecasting wind speed, reconstructs windy fast data linear superposition to obtain wind speed value.
2. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 1, it is characterised in that:
The initial value and threshold value of Genetic Algorithm Optimized Neural Network are utilized in step 3；Specific step is as follows:
Step 1: random initializtion population: individual UVR exposure uses real coding mode, and each individual is a real number string, should
Real number string includes input layer and hidden layer connection weight, hidden layer threshold value, hidden layer and output layer connection weight and output layer
Threshold value；
Step 2: fitness function is determined: with falling for the absolute error sum between the prediction output of neural network and desired output
Number is used as fitness function F；
Step 3: selection operation: selecting several body as parents for raising up seed from population, and the high individual of fitness is lost
Pass to that follow-on probability is larger, the low individual inheritance of fitness is then smaller to follow-on probability；
Step 4: crossover operation: the individual of two pairings is with crossover probability pcPart of gene is exchanged, two new are formed
Body；
Step 5: mutation operation: with a smaller mutation probability pvJ-th of gene g of i-th of individual of selectionijBecome
It is different；
Step 6: fitness function value is calculated, the weight and threshold value of optimization are exported if meeting algorithm termination condition, if being unsatisfactory for
Algorithm termination condition then return step three；
Step 7: using the weight of the optimization of genetic algorithm output and threshold value as the initial weight and threshold value of neural network, with instruction
Practice sample to neural metwork training, obtains short-term wind speed forecasting model.
3. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 1, it is characterised in that:
Using the penalty factor and kernel function ginseng in the wind speed combination forecasting of particle swarm algorithm Support Vector Machines Optimized in step 3
Optimal penalty factor and kernel functional parameter g are assigned to support vector machines by number g, and with the wind speed sample of building to supporting vector
Machine network training establishes the wind speed combination forecasting of optimization, analysis and evaluation and foreca result.
4. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 3, it is characterised in that:
Using the penalty factor and kernel functional parameter g in the wind speed combination forecasting of particle swarm algorithm Support Vector Machines Optimized
Specific practice be:
A group particle is initialized in solution space, each particle I represents the penalty factor and kernel functional parameter g of built-up pattern
A potential optimal solution, with position vector XI=(xI1,xI2,…xIn) velocity vector VI=(vI1,vI2,…vIn) two n tie up to
Amount is to indicate, speed and location updating equation are as follows:
In formula, Indicate the particle position and speed that d is tieed up in k iteration Chinese style；W indicates weight；c1, c2It indicates to learn
The factor is practised, Chang Jun takes 2,Take the random number between [0,1]；Indicate particle I individual extreme value the
The coordinate of d dimension；Indicate that entire group's global extreme point searches fitness value in the coordinate that d is tieed up after successive ignition
Optimal location, as optimal penalty factor and kernel functional parameter g.
5. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 1, feature
Be: the number of plies n of signal component is equal to 5 in step 2.
6. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 5, feature
Be: five layer signal components include one layer of low frequency signal and four layers of high-frequency signal.
7. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 4, feature are being weighed
Be conducive to: the kernel function that step 3 Kernel Function parameter g is used is Gaussian radial basis function K (xi,xj)=exp [- (| | xi-xj|
|2)/σ2]；X in formulaiIt is spatially any point, xjIt is kernel function center, σ is the width parameter of function, and the radial of control function makees
Use range.
8. a kind of wind farm wind velocity combination forecasting method based on wavelet analysis according to claim 1, feature
It is: before the data sample in step 1 does normalization operation, also the exceptional value of acquired air speed data is carried out at tranquilization
Reason studies for a second time courses one has flunked missing data, chooses data sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811546166.8A CN109615142A (en) | 2018-12-18 | 2018-12-18 | A kind of wind farm wind velocity combination forecasting method based on wavelet analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811546166.8A CN109615142A (en) | 2018-12-18 | 2018-12-18 | A kind of wind farm wind velocity combination forecasting method based on wavelet analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109615142A true CN109615142A (en) | 2019-04-12 |
Family
ID=66009440
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811546166.8A Pending CN109615142A (en) | 2018-12-18 | 2018-12-18 | A kind of wind farm wind velocity combination forecasting method based on wavelet analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109615142A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506868A (en) * | 2020-04-07 | 2020-08-07 | 河海大学 | Ultrashort-term wind speed prediction method based on HHT weight optimization |
| CN113762602A (en) * | 2021-08-13 | 2021-12-07 | 中国大唐集团科学技术研究院有限公司西北电力试验研究院 | Short-term wind speed prediction method for wind power plant |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063641A (en) * | 2010-10-14 | 2011-05-18 | 北京大学 | Method for forecasting wind speed of high speed railway line |
| CN102542133A (en) * | 2010-12-10 | 2012-07-04 | 中国科学院深圳先进技术研究院 | Short-time wind speed forecasting method and system for wind power plant |
| CN105184391A (en) * | 2015-08-19 | 2015-12-23 | 国网山东省电力公司电力科学研究院 | Method for predicting wind speed and power of wind farm based on wavelet decomposition and support vector machine |
-
2018
- 2018-12-18 CN CN201811546166.8A patent/CN109615142A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063641A (en) * | 2010-10-14 | 2011-05-18 | 北京大学 | Method for forecasting wind speed of high speed railway line |
| CN102542133A (en) * | 2010-12-10 | 2012-07-04 | 中国科学院深圳先进技术研究院 | Short-time wind speed forecasting method and system for wind power plant |
| CN105184391A (en) * | 2015-08-19 | 2015-12-23 | 国网山东省电力公司电力科学研究院 | Method for predicting wind speed and power of wind farm based on wavelet decomposition and support vector machine |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506868A (en) * | 2020-04-07 | 2020-08-07 | 河海大学 | Ultrashort-term wind speed prediction method based on HHT weight optimization |
| CN111506868B (en) * | 2020-04-07 | 2023-08-25 | 河海大学 | Ultra-short-term wind speed prediction method based on HHT weight optimization |
| CN113762602A (en) * | 2021-08-13 | 2021-12-07 | 中国大唐集团科学技术研究院有限公司西北电力试验研究院 | Short-term wind speed prediction method for wind power plant |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112508275B (en) | Power distribution network line load prediction method and equipment based on clustering and trend indexes | |
| CN109711620B (en) | Short-term power load prediction method based on GRU neural network and transfer learning | |
| Ghadimi et al. | PSO based fuzzy stochastic long-term model for deployment of distributed energy resources in distribution systems with several objectives | |
| CN108205717A (en) | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology | |
| CN109146162B (en) | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network | |
| CN103268366A (en) | Combined wind power prediction method suitable for distributed wind power plant | |
| CN110674999A (en) | Cell load prediction method based on improved clustering and long-short term memory deep learning | |
| CN111753893A (en) | Wind turbine generator power cluster prediction method based on clustering and deep learning | |
| CN105046374A (en) | Power interval predication method based on nucleus limit learning machine model | |
| CN105512745A (en) | Wind power section prediction method based on particle swarm-BP neural network | |
| CN104636985A (en) | Method for predicting radio disturbance of electric transmission line by using improved BP (back propagation) neural network | |
| CN105069236B (en) | Consider the broad sense load joint probability modeling method of wind power plant node space correlation | |
| CN108428017A (en) | Wind power interval prediction method based on core extreme learning machine quantile estimate | |
| CN109255726A (en) | A kind of ultra-short term wind power prediction method of Hybrid Intelligent Technology | |
| CN110264012A (en) | Renewable energy power combination prediction technique and system based on empirical mode decomposition | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| Hu et al. | Short-term photovoltaic power prediction based on similar days and improved SOA-DBN model | |
| CN109858665A (en) | Photovoltaic short term power prediction technique based on Feature Selection and ANFIS-PSO | |
| CN109242136A (en) | A kind of micro-capacitance sensor wind power Chaos-Genetic-BP neural network prediction technique | |
| CN110163444A (en) | A kind of water demand prediction method based on GASA-SVR | |
| CN102184328A (en) | Method for optimizing land use evolution CA model transformation rules | |
| Chen et al. | Research on wind power prediction method based on convolutional neural network and genetic algorithm | |
| CN116826710A (en) | Peak clipping strategy recommendation method and device based on load prediction and storage medium | |
| CN109615142A (en) | A kind of wind farm wind velocity combination forecasting method based on wavelet analysis | |
| CN113344283B (en) | Energy internet new energy consumption capability assessment method based on edge intelligence |
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: 20190412 |