CN108154260A - A kind of short-term wind power forecast method - Google Patents
A kind of short-term wind power forecast method Download PDFInfo
- Publication number
- CN108154260A CN108154260A CN201711360489.3A CN201711360489A CN108154260A CN 108154260 A CN108154260 A CN 108154260A CN 201711360489 A CN201711360489 A CN 201711360489A CN 108154260 A CN108154260 A CN 108154260A
- Authority
- CN
- China
- Prior art keywords
- climbing
- wind
- data
- prediction
- wind speed
- 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 47
- 238000005070 sampling Methods 0.000 claims abstract description 14
- 230000009194 climbing Effects 0.000 claims description 49
- 238000009826 distribution Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 241001269238 Data Species 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 2
- 239000011159 matrix material Substances 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 14
- 238000002474 experimental method Methods 0.000 description 7
- 238000013461 design Methods 0.000 description 5
- 238000000926 separation method Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000004940 physical analysis method Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 230000001052 transient 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
- G06Q10/04—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
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)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Wind Motors (AREA)
Abstract
Feature and the short-term wind power forecast method of random sampling combination on-line study machine are lifted based on wind speed the invention discloses a kind of, the method is：Wind speed lifting feature is added in as mark value in original wind series and carries out wind power prediction, the wind series for adding in characteristic value are sampled by random sampling combined method and are combined into multiple new sample sets, then on-line study machine model is established to each new sample set and carries out short-term wind-electricity power prediction.The present invention has the characteristics of generalization ability is strong, and precision of prediction is high in short-term wind-electricity prediction, has certain practical value.
Description
Technical field
The present invention relates to wind power prediction field, specially a kind of short-term wind power forecast method.
Background technology
With the swift and violent increase of installed capacity of wind-driven power, the intrinsic fluctuation of wind and randomness become the bottle of China's Wind Power Development
Neck, large-scale wind-electricity integration propose higher requirement to wind power prediction.
Nowadays wind power forecasting method is roughly divided into physical analysis method, statistical learning method and the two and combines.Physics side
The physical process that method changes suitable for wind in addition wind power plant geography information and blower fan group characteristic are analyzed, modeled and predicted, is formed
Blower fan group or wind power curve carry out prediction power.Statistical learning method is based on historical data, such as NWP data, wind-force
Generator and anemometer tower gathered data etc. by trend analysis, obtain the model that wind converts to power to be predicted, should
Requirement of the method to historical data is higher.Common statistical learning method has neural network, support vector machines, wavelet analysis
Deng, with the development of wind power prediction technology, researchers also more improve the attention degree that wind speed rapidly rises or falls, claim
For " wind speed climbing ", if wind speed climbing is added in wind power prediction, precision of prediction will necessarily be improved.
Though traditional neural network such as BP neural network has stronger nonlinear Identification ability, generalization ability is insufficient,
Pace of learning becomes slowly restricts the reason of it develops；Extreme learning machine (Extreme Learning Machine, ELM), with
Simple in structure, pace of learning is fast and generalization ability is well advantage, and in electric system application, ELM is pre- in real-time wind power
It surveys, electric power system transient stability online evaluation etc. has higher rating.And there may be offices for the prediction error of single model
The larger possibility in portion, the efficient combination of multiple models can share the error of each model so that prediction result more has generation
Table.To improve the efficiency and precision of super short-period wind power prediction, propose based on Bootstrap (random sampling combination) and
The short-term wind-electricity power prediction model of OS-ELM (on-line study machine model), although prediction effect is improved, model prediction
As a result efficient combination is not obtained, Bootstrap can combine multiple time-dependent models so that prediction result is more stablized.It is another
Aspect, the lifting situation of wind speed is very important, and the accuracy of wind power prediction can be effectively improved by adding in the lifting feature of wind speed.
Invention content
The purpose of the present invention is to solve defects in the prior art, provide a kind of pre- for short-term wind-electricity power
The method of survey is to lift feature and the short-term wind power forecast method of Bootstrap-ELM based on wind speed.This method uses root
It is characterized as that original wind series increase mark value and add in wind speed lifting feature in wind power prediction according to wind speed, then passes through
Bootstrap-OS-ELM models predict power.
To achieve the above object, the present invention provides following technical solution：
The short-term wind power forecast method, includes the following steps：
Step 1) acquires the measured data (v (t), p (t)) of wind power plant, and wherein t is the sampling time, and is the nature more than 0
Number, v (t) are actual measurement wind speed, and p (t) is measured power；
Step 2) carries out data prediction with reference to periphery wind field data and historical data, and the data prediction includes picking
Except exceptional value always fills up missing data, obtaining the data after denoising is Respectively
Wind speed and power data after denoising；
Step 3) introduces wind speed lifting feature：If i and j moment wind speed is respectively viAnd vj, vrampThe width of climbing for setting
Degree, VrampFor the practical gradient of climbing, time interval Δ T=| j-i |, when the wind speed gradient of climbing in time interval Δ T reaches
vramp, then it is assumed that climbing event has occurred in wind speed, and sets climbing type；
Step 4) is sampled recombinant using Bagging methods to the wind series for adding in climbing characteristic：First with
Bootstrap methods are sampled original sample with repeatedly putting back to, and are combined into multiple new sample sets, then to each
Sample set is established regression model respectively, is finally obtained the prediction result identical with new samples collection quantity, is passed through traditional system
All prediction results are combined by meter method or machine learning method；
Step 5) establishes Bagging-OS-ELM models and carries out short-term wind-electricity power prediction.
The further design of the short-term wind power forecast method is, in the step 3), climbing type is pros
Climb to climbing and negative direction.
The further design of the short-term wind power forecast method is, in the step 3), definition climbing type tool
Body is：In defined time interval Δ T, if v (t+ Δ T) ＜ v (t) andThen in this Δ T time
Wind speed power data be defined as positive direction climbing；If v (t+ Δ T) ＞ v (t) andThen during this Δ T
Interior wind speed power data is defined as the second class, is defined as negative direction climbing.
The further design of the short-term wind power forecast method is, in the step 3), if v (t+ Δ T)=v
(t) orThen ignore influence of the variation to generated output of wind speed in time period, be defined as not sending out
Raw climbing event.
The further design of the short-term wind power forecast method is that the detailed process of the step 4) is：Step
4-1) assume Xi=xi, Xi~i.i.d.F, i=1,2 ..., n obey F distributions, make X=(X1,X2,…,Xn) and x=(x1,
x2,…,xn), respectively random sample and sample observations, it is assumed that stochastic variable θ=S (X), based on observation x, estimation
Stochastic variable θ is distributed；
Step 4-2) distribution of construction X experiencesPut back to fromN subsample of middle extraction, it is as follows：
It enablesWithWithSelf-service distribution estimate θ=S
(X) sample distribution.
Step 4-3) it is extracted in set X with putting back to and obtains B self-service sample X*1, X*2... X*B, wherein all samples
Independently of each other, B values are among 25 and 200, for each X*b,
Step 4-4) to standard deviationEstimation beStandard deviation, wherein
The further design of the short-term wind power forecast method is that the step 5) specifically includes：Initialize rank
Section：The initial node in hidden layer of given network, inputs initial training sample data, initializes network and sets network parameter,
Acquire initial hidden layer output matrix H0With output weight vector β0；The online sequence more new stage：On the initial network basis of foundation
On, according to newest sample data undated parameter H and β, until all sample datas learn to complete.
Beneficial effects of the present invention
(1) the Bagging-OS-ELM methods OS-ELM method stability that compares is stronger, and predicts that error is relatively small, more
It is suitble to the short term power prediction of wind power plant；
(2) it is compared with 0S-ELM, F-OS-ELM and Bagging-OS-ELM, feature and Bootstrap- is lifted based on wind speed
The root-mean-square error and absolute average error of the method for ELM are minimum, more preferable to the fitting effect of actual power.
Description of the drawings
Fig. 1 is Bootstrap schematic diagrams.
Fig. 2 is Bagging-OS-ELM flow charts.
The performance that Fig. 3 is OS-ELM and Bagging-OS-ELM compares.
Fig. 4 is to determine Bootstrap double sampling number figures.
Fig. 5 is the January of tetra- kinds of models of 0S-ELM, F-OS-ELM, Bagging-OS-ELM and F-Bagging-OS-ELM
Prediction effect.
Fig. 6 is the March of tetra- kinds of models of 0S-ELM, F-OS-ELM, Bagging-OS-ELM and F-Bagging-OS-ELM
Prediction effect.
Fig. 7 is the September of tetra- kinds of models of 0S-ELM, F-OS-ELM, Bagging-OS-ELM and F-Bagging-OS-ELM
Prediction effect.
Fig. 8 is the December of tetra- kinds of models of 0S-ELM, F-OS-ELM, Bagging-OS-ELM and F-Bagging-OS-ELM
Prediction effect.
Fig. 9 is the absolute error schematic diagram of positive climbing.
Figure 10 is the absolute error schematic diagram of negative sense climbing.
Description of the drawings
The technical solution of invention is described in detail below in conjunction with the accompanying drawings.
The present invention's is included the following steps based on wind speed lifting feature and random sampling combination on-line study machine：
Step 1) acquires the measured data (v (t), p (t)) of wind power plant, and wherein t is the sampling time, and is the nature more than 0
Number, v (t) are actual measurement wind speed, and p (t) is measured power.
Step 2) carries out data prediction with reference to periphery wind field data and historical data, and the data prediction includes picking
Except exceptional value always fills up missing data, obtaining the data after denoising is Respectively
Wind speed and power data after denoising.
Step 3) introduces wind speed lifting characterization method, adds mark value for wind farm wind velocity data, detailed process is：
Step 3-1) to set i and j moment wind speed be respectively viAnd vj, vrampFor the amplitude of climbing, i.e. threshold value, VrampFor reality
The climbing gradient, formula are：Time interval Δ T=j-i.
Step 3-2) in defined time interval Δ T, if v (t+ Δ T) ＜ v (t) andThen this
Wind speed power data in Δ T time is defined as the first kind, climbs for positive direction.If v (t+ Δ T) ＞ v (t) andThen the wind speed power data in this Δ T time is defined as the second class, climbs for negative direction.If v (t+
Δ T)=v (t) orThen the wind speed power data in this Δ T time is defined as that climbing event does not occur,
The climbing gradient is ignored.
Step 4) constructs self-service sample by resampling using Bootstrap methods, small to not belonging to solve sample number
In the challenge of normal distribution, basic thought is：
Step 4-1) assume Xi=xi, Xi~i.i.d.F, i=1,2 ..., n.F distributions are obeyed, make X=(X1,X2,…,Xn)
With x=(x1,x2,…,xn), respectively random sample and sample observations.Assuming that stochastic variable θ=S (X), using observation x as
Basis, the θ distributions of estimation stochastic variable.
Step 4-2) construction X experiences distribution firstHave put back to fromN subsample of middle extraction, it is as follows
It enablesWithWithSelf-service distribution estimate θ=S
(X) sample distribution.
Step 4-3) and then there is the extraction in set X put back to obtain B self-service sample X*1,X*2,…X*B, wherein all
Sample is mutual indepedent.The general values of B are among 25 and 200, and B takes 200 in the present embodiment.For each X*b,
Step 4-4) finally to standard deviationMust estimate beStandard deviation, wherein
A specific example presented below and experimental data, and specifically illustrated with reference to the above method：
Data source is randomly selected in the wind speed power data of Shanghai wind power plant 2014-2015 in 1 year 2015
5000 groups of data, wherein 80% be used as training data, 20% be used as test data.Case verification is specially：
1, verify Bagging-OS-ELM superiority
Respectively 100 experiments, record are carried out using single OS-ELM models and Bagging-OS-ELM weights built-up pattern
The predicted root mean square error RMSE of lower two models, 100 groups of experiments.
As shown in figure 3,100 result of the tests of as two models, as seen from the figure, Bagging-OS-ELM models it is pre-
Survey error it is relatively stable, Stable distritation between 81-81.2, and OS-ELM models in 100 groups of experiments distribution of results range compared with
Greatly, about between 80.4 to 82.2, it can be seen that the Bagging-OS-ELM methods OS-ELM method stability that compares is stronger,
And prediction error is relatively small, so using the prediction model of Bagging weight combined strategies, is more suitable for the short-term work(of wind power plant
Rate is predicted.
2, determine bootstrap frequency in samplings
In Bagging weight combined strategies, the prediction result of prediction model and Bootstrap frequency in sampling n correlations compared with
Greatly, n is the frequency in sampling having in the slave original sample put back to, to determine optimal sampling number, respectively with Bagging-OS-
ELM models be sampled number from 1 to 200 time in the case of 200 groups of experiments, such as scheme shown in (4), frequency in sampling is bigger, model
It tends towards stability, when frequency in sampling is more than 80, the prediction effect and stability of model are preferable, so this chapter is with n=80 progress
Further experiment.
3, determine time interval
It keeps presetting parameter optimal value constant, chooses different time intervals and carry out contrast test.As shown in Table 1, when
When Δ T is 10min, the separation prediction of wind speed power data can obtain preferable test effect.For from theory significance, with compared with
Short Δ T can more completely detach original signal, can reduce the presence of residual component, if Δ T is larger, 1h or 2h are longer
Period in wind speed uncertainty may generate several times just climbing or negative climbing event, and 10min in occur it is such
Situation it is relatively low.The purpose of separation initial data is by positive and negative climbing and the separation of event maximum possible of not climbing, if Δ
T values are larger or excessively small, such as several seconds to a few minutes, then separation will become meaningless.
1 time interval Δ T values of table
4, verify F-Bagging-OS-ELM model superiority
March is respectively adopted in 2015 in this experiment, four groups of June, September and December data are tested, and data are complete
Year anemometer tower and wind turbine gathered data, training data of the data of 29 days as model before every group of data, last day is as survey
Data are tried, wind speed power data temporal resolution is 5min, for the superiority of verification method, compared 0S-ELM, F- respectively
The prediction result of tetra- kinds of predictions of OS-ELM, Bagging-OS-ELM and F-Bagging-OS-ELM, every group of data carry out four kinds of moulds
Type is tested, accurate come the prediction of evaluation model by root-mean-square error (RMSE), absolute average error (MAE) and absolute error
Degree.
As shown in table 2, situation is compared for the prediction RMSE of the lower four kinds of models of four groups of experimental datas, as seen from table F-
The precision of prediction of Bagging-OS-ELM is better than its excess-three kind, and OS-ELM models are poor with respect to prediction effect, and Bagging-OS-
Slightly better than wind speed lifting feature OS-ELM models are added in, F-Bagging-OS-ELM is combined the prediction effect of ELM built-up patterns
The advantage of the two methods.From the time, in the exactly winter in summer in June and December, fluctuations in wind speed is larger, prediction result fluctuation
It is larger, but dry monsoon is relatively large, thus the RMSE in December is integrally larger, and M ＆ S is relatively stablized, the entirety in March
RMSE is relatively small.But the lower four groups of predictions RMSE mean values of the Same Way that compares, this chapter, which is carried, adds in wind speed lifting feature
Bagging weight combined methods are more superior.Table 3 compares situation for four kinds of model prediction result MAE, and the effect of MAE is evaluation
The average amplitude of error.
2 each model prediction result (RMSE) of table
Tab.2 Different methods of power prediction(RMSE)
3 each model prediction result (MAE) of table
Tab.3 Different methods of power prediction(MAE)
If Fig. 5 to Fig. 8 is to put forward the wind power prediction value of four kinds of models of the previous day and pair of actual power value in four middle of the month
Than figure.It is good to the fitting effect of actual power to make a general survey of four figures, the September that compares and December, the fitting degree in January and March
More preferably, by observing four kinds of model prediction broken lines and actual power broken line, it is possible to find the fitting effect of F-Bagging-OS-ELM
Relatively preferably.
5) the negative climbing prediction effect comparison of positive climbing
Such as Fig. 9, Figure 10, respectively correspondingly predicted for positive, negative sense climbing absolute to the prediction of following 24 time intervals
Error Graph, RMSE of the table 4 for positive and negative climbing prediction result.Prediction model carries built-up pattern by one's own department or unit.With 2015 6,9 and 12
Month the whole month wind speed power data carries out 50 groups of experiments respectively to just climbing and bear climbing situation and consensus forecast is as a result, when climbing amount
In 0.4M/S, when the climbing time is 10 minutes, the prediction result of contrast table 4 can be seen that prediction algorithm is climbed for just climbing and bearing
The prediction result on slope differs greatly, and illustrates to influence wind power prediction model precision when positive climbing and negative climbing occur for wind speed
It is larger, while can be seen that when the wind power prediction result when the positive climbing of wind speed generation is climbed relative to wind speed generation is negative
Precision is poor, illustrates that influence of the wind speed feature difference to model prediction result is different.
4 positive and negative climbing prediction result of table
Claims (6)
1. a kind of short-term wind power forecast method, it is characterised in that include the following steps：
Step 1) acquires the measured data (v (t), p (t)) of wind power plant, and wherein t is the sampling time, and is natural number more than 0, v
(t) it is actual measurement wind speed, p (t) is measured power；
Step 2) carries out data prediction with reference to periphery wind field data and historical data, and the data prediction includes rejecting different
Constant value always fills up missing data, obtains the data after denoising and is Respectively denoising
Wind speed and power data afterwards；
Step 3) introduces wind speed lifting feature：If i and j moment wind speed is respectively viAnd vj, vrampThe amplitude of climbing for setting,
VrampFor the practical gradient of climbing, time interval Δ T=| j-i |, when the wind speed gradient of climbing in time interval Δ T reaches vramp, then
Think that climbing event has occurred in wind speed, and set climbing type；
Step 4) is sampled recombinant using Bagging methods to the wind series for adding in climbing characteristic：First with
Bootstrap methods are sampled original sample with repeatedly putting back to, and are combined into multiple new sample sets, then to each
Sample set is established regression model respectively, is finally obtained the prediction result identical with new samples collection quantity, is passed through traditional system
All prediction results are combined by meter method or machine learning method；
Step 5) establishes Bagging-OS-ELM models and carries out short-term wind-electricity power prediction.
2. short-term wind power forecast method according to claim 1, it is characterised in that in the step 3), type of climbing
Climb for positive direction climbing with negative direction.
3. short-term wind power forecast method according to claim 2, it is characterised in that in the step 3), definition climbing
Type is specially：In defined time interval Δ T, if v (t+ Δ T) ＜ v (t) andThen this Δ T
Wind speed power data in time is defined as positive direction climbing；If v (t+ Δ T) ＞ v (t) andThen this
Wind speed power data in Δ T time is defined as the second class, is defined as negative direction climbing.
4. short-term wind power forecast method according to claim 1, it is characterised in that in the step 3), if v (t+ Δs
T)=v (t) orThen ignore influence of the variation to generated output of wind speed in time period, define
For climbing event does not occur.
5. short-term wind power forecast method according to claim 1, it is characterised in that the step 4) specifically includes
Following steps：
Step 4-1) assume Xi=xi, Xi~i.i.d.F, i=1,2 ..., n obey F distributions, make X=(X1,X2,…,Xn) and x
=(x1,x2,…,xn), respectively random sample and sample observations, it is assumed that stochastic variable θ=S (X), using observation x as base
Plinth, the θ distributions of estimation stochastic variable；
Step 4-2) distribution of construction X experiencesPut back to fromN subsample of middle extraction, it is as follows：
It enablesWithWithSelf-service distribution estimate θ=S (X) sample
This distribution.
Step 4-3) it is extracted in set X with putting back to and obtains B self-service sample X*1, X*2... X*B, wherein all samples are mutual
Independent, B values are among 25 and 200, for each X*b,B=1,2 ..., B；
Step 4-4) to standard deviationEstimation beStandard deviation, wherein
6. short-term wind power forecast method according to claim 1, it is characterised in that the step 5) specifically includes：
Initial phase：The initial node in hidden layer of given network, inputs initial training sample data, initializes network and sets
Network parameter is put, acquires initial hidden layer output matrix H0With output weight vector β0；
The online sequence more new stage：On the basis of the initial network of foundation, according to newest sample data undated parameter H and β, directly
Learn to complete to all sample datas.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711360489.3A CN108154260A (en) | 2017-12-15 | 2017-12-15 | A kind of short-term wind power forecast method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711360489.3A CN108154260A (en) | 2017-12-15 | 2017-12-15 | A kind of short-term wind power forecast method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108154260A true CN108154260A (en) | 2018-06-12 |
Family
ID=62467366
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711360489.3A Pending CN108154260A (en) | 2017-12-15 | 2017-12-15 | A kind of short-term wind power forecast method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108154260A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109494716A (en) * | 2018-11-15 | 2019-03-19 | 沈阳工业大学 | Wind power output power confidence interval prediction technique based on Bootstrap |
| CN109725627A (en) * | 2018-12-06 | 2019-05-07 | 南京信息工程大学 | A kind of wind power climbing event detection system and detection method |
| CN110543929A (en) * | 2019-08-29 | 2019-12-06 | 华北电力大学（保定） | wind speed interval prediction method and system based on Lorenz system |
| CN111539577A (en) * | 2020-04-29 | 2020-08-14 | 南京信息工程大学 | Short-term wind power prediction method based on wind speed change rate and Gaussian process regression |
| CN112485394A (en) * | 2020-11-10 | 2021-03-12 | 浙江大学 | Water quality soft measurement method based on sparse self-coding and extreme learning machine |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106374465A (en) * | 2016-11-10 | 2017-02-01 | 南京信息工程大学 | GSA-LSSVM model-based short period wind electricity generation power prediction method |
| CN106933778A (en) * | 2017-01-22 | 2017-07-07 | 中国农业大学 | A kind of wind power combination forecasting method based on climbing affair character identification |
-
2017
- 2017-12-15 CN CN201711360489.3A patent/CN108154260A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106374465A (en) * | 2016-11-10 | 2017-02-01 | 南京信息工程大学 | GSA-LSSVM model-based short period wind electricity generation power prediction method |
| CN106933778A (en) * | 2017-01-22 | 2017-07-07 | 中国农业大学 | A kind of wind power combination forecasting method based on climbing affair character identification |
Non-Patent Citations (1)
| Title |
|---|
| 敖培等: "基于混合扰动的超短期风电功率ELM集成预测", 《数字技术与应用》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109494716A (en) * | 2018-11-15 | 2019-03-19 | 沈阳工业大学 | Wind power output power confidence interval prediction technique based on Bootstrap |
| CN109494716B (en) * | 2018-11-15 | 2022-04-26 | 沈阳工业大学 | Bootstrap-based wind power output power confidence interval prediction method |
| CN109725627A (en) * | 2018-12-06 | 2019-05-07 | 南京信息工程大学 | A kind of wind power climbing event detection system and detection method |
| CN110543929A (en) * | 2019-08-29 | 2019-12-06 | 华北电力大学（保定） | wind speed interval prediction method and system based on Lorenz system |
| CN110543929B (en) * | 2019-08-29 | 2023-11-14 | 华北电力大学（保定） | Wind speed interval prediction method and system based on Lorenz system |
| CN111539577A (en) * | 2020-04-29 | 2020-08-14 | 南京信息工程大学 | Short-term wind power prediction method based on wind speed change rate and Gaussian process regression |
| CN112485394A (en) * | 2020-11-10 | 2021-03-12 | 浙江大学 | Water quality soft measurement method based on sparse self-coding and extreme learning machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108154260A (en) | A kind of short-term wind power forecast method | |
| CN105046374B (en) | A kind of power interval prediction technique based on core extreme learning machine model | |
| Wang et al. | Comparison of seven methods for determining the optimal statistical distribution parameters: A case study of wind energy assessment in the large-scale wind farms of China | |
| CN109063276B (en) | Wind power plant dynamic equivalent modeling method suitable for long-time domain wind speed fluctuation | |
| CN111753893A (en) | Wind turbine generator power cluster prediction method based on clustering and deep learning | |
| CN104504508B (en) | The taiwan area closed-loop data analysis method returned based on step analysis and small echo | |
| CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
| CN104156889B (en) | A kind of wind power plant performance evaluation system and its appraisal procedure based on WAMS data | |
| CN105303250A (en) | Wind power combination prediction method based on optimal weight coefficient | |
| Suomalainen et al. | Synthetic wind speed scenarios including diurnal effects: Implications for wind power dimensioning | |
| CN111475909B (en) | Wind turbine generator output correlation mapping modeling method based on long-term and short-term memory network | |
| CN103996079B (en) | Wind power weighting predication method based on conditional probability | |
| CN105574615A (en) | Spatial correlation and genetic algorithm (GA) based wind power forecast method for wavelet-BP neural network | |
| CN103400009A (en) | Wind electric field dynamic equivalence method based on split level semi-supervised spectral clustering algorithm | |
| CN105825002B (en) | A kind of wind power plant dynamic equivalent modeling method based on dynamic Gray Association Analysis | |
| CN103870923A (en) | Information entropy condensation type hierarchical clustering algorithm-based wind power plant cluster aggregation method | |
| CN109255728A (en) | The photovoltaic generation power neural network prediction method of chaos phase space optimal reconfiguration | |
| CN108717579A (en) | A kind of short-term wind-electricity power interval prediction method | |
| CN105741192B (en) | Short-term wind speed combined forecasting method for wind turbine engine room of wind power plant | |
| Cheggaga et al. | A neural network solution for extrapolation of wind speeds at heights ranging for improving the estimation of wind producible | |
| Chikuni et al. | Estimating wind power generation capacity in Zimbabwe using vertical wind profile extrapolation techniques: A case study | |
| Zheng et al. | Characteristics for wind energy and wind turbines by considering vertical wind shear | |
| CN113034018B (en) | Power grid supply and demand balance adjustment method based on power moment analysis | |
| CN112290538A (en) | Load model parameter online correction method based on aggregation-identification double-layer framework | |
| CN113224748A (en) | Method for calculating line loss of low-voltage distribution station area |
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: 20180612 |