CN108898251A - Consider the marine wind electric field power forecasting method of meteorological similitude and power swing - Google Patents
Consider the marine wind electric field power forecasting method of meteorological similitude and power swing Download PDFInfo
- Publication number
- CN108898251A CN108898251A CN201810696196.0A CN201810696196A CN108898251A CN 108898251 A CN108898251 A CN 108898251A CN 201810696196 A CN201810696196 A CN 201810696196A CN 108898251 A CN108898251 A CN 108898251A
- Authority
- CN
- China
- Prior art keywords
- power
- wind
- power swing
- similitude
- meteorological
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000005684 electric field Effects 0.000 title claims abstract description 18
- 238000013277 forecasting method Methods 0.000 title claims abstract description 8
- 238000013528 artificial neural network Methods 0.000 claims abstract description 14
- 238000005096 rolling process Methods 0.000 claims abstract description 10
- 238000012937 correction Methods 0.000 claims abstract description 9
- 238000000034 method Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 11
- 238000000513 principal component analysis Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000000540 analysis of variance Methods 0.000 claims description 2
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 230000002123 temporal effect Effects 0.000 claims description 2
- 238000007405 data analysis Methods 0.000 claims 1
- 230000009194 climbing Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 238000012360 testing method Methods 0.000 description 5
- 230000005611 electricity Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000000053 physical method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 101100129500 Caenorhabditis elegans max-2 gene Proteins 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000003860 storage 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- H02J3/386—
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/76—Power conversion electric or electronic aspects
-
- 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
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Power Engineering (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Wind Motors (AREA)
Abstract
The present invention relates to a kind of marine wind electric field power forecasting methods for considering meteorological similitude and power swing, include the following steps:1) the wind energy data in marine wind electric field certain time are obtained, and as original wind energy sample;2) clustering that meteorological similitude is carried out to original wind energy sample data, obtains meteorological similitude classification results;3) classify to power swing, and obtain the generic of power swing range using extreme learning machine ELM, obtain the classification results of power swing;4) using tentative prediction model of the Elman neural network based on sample similarity, the short-term wind speed rolling forecast of wind power plant is carried out to future time instance, and obtains initial predicted result;5) NWP correction model is established using multilayer perceptron MLP to be modified initial predicted result, obtain revised final result.Compared with prior art, the present invention has many advantages, such as to consider that meteorological similitude and power swing, prediction are accurate comprehensive.
Description
Technical field
The present invention relates to marine power prediction technical fields, more particularly, to a kind of meteorological similitude of consideration and power swing
Marine wind electric field power forecasting method.
Background technique
Offshore wind farm is since wind-force is stablized, and close to load center, being not take up the reasons such as land resource becomes the following wind-powered electricity generation row
The prior development direction of industry.However as the increase of installed capacity, the fluctuation and randomness of offshore wind farm power output are to system
Stable operation brings lot of challenges.Therefore electric system is scheduled by the prediction of accurate short term power, for electric power
System stable operation and increasing economic efficiency is of great significance.Traditional wind power prediction method can be mainly divided into wind
The physical method of the profile row detailed description of electric field whole region;And by analysis mass data, find weather information and wind
The statistical method that potential statistical relationship is predicted between electric field electricity-generating amount.Either physical model or statistical model, number
It is worth weather forecast information (NWP) information usually as one group of important input data of above-mentioned prediction model.This is because for
Duration is more than prediction in three hours, and NWP data reflect the essence of air motion.With regard to physical method level, due to marine environment
The reasons such as complicated, wake effect range is wide, climatological data shortage cause marine physical modeling operand cumbersome, and modeling is difficult;And
And due to the difference of geographical environment, the flexible generalization of physical model is bad.And the Extrapolating model method based on Statistics can make
It is trained with given meteorological condition, avoids the physical modeling step of intermediate complexity;But on the other hand, China sea NWP
The low precision of information and the significantly variation of wind power output directly affect the accuracy of statistical model.
Summary of the invention
It is meteorological similar that it is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide a kind of considerations
The marine wind electric field power forecasting method of property and power swing.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of marine wind electric field power forecasting method considering meteorological similitude and power swing, includes the following steps:
1) the wind energy data in marine wind electric field certain time are obtained, and as original wind energy sample;
2) Principal Component Analysis and self organizing neural network is combined to carry out meteorological similitude to original wind energy sample data
Clustering obtains meteorological similitude classification results;
3) classify to power swing, and obtain the generic of power swing range using extreme learning machine ELM, obtain
To the classification results of power swing;
4) using tentative prediction model of the Elman neural network based on sample similarity, according to meteorological similitude
The classification results of classification results and power swing choose training sample relevant to prediction time and carry out wind power plant to future time instance
Short-term wind speed rolling forecast, and obtain initial predicted result;
5) NWP correction model is established using multilayer perceptron MLP to be modified initial predicted result, obtain revised
Final result.
Preferably, in the step 1), original wind energy sample is expressed as S in the vector space of t momentt, data knot
Structure is:
Wherein, WS, WP, WD, T, W, P respectively indicate wind speed, power, wind direction, temperature, humidity and barometric information, subscript
Indicate value at the time of data, WStFor the wind speed forecasting data of prediction time t, WSt-T1For the history wind at T1 moment before prediction time t
Speed value.
Preferably, the step 2) specifically includes following steps:
21) principal component is extracted by Principal Component Analysis, and obtains the contribution rate of corresponding principal component by cross-correlation matrix,
And choosing cumulative proportion in ANOVA is more than 90% corresponding principal component;
22) using the principal component after screening as the input variable of self organizing neural network, cluster centre quantity is set, is carried out
Similitude clusters to obtain preliminary meteorological similitude classification.
Preferably, the step 3) specifically includes following steps:
31) power swing is fallen into 5 types by power swing amplitude, specially:
The first kind:Power swing in -15%Pr hereinafter,
Second class:Power swing in -15%Pr~0,
Third class:Power swing in 0~10%Pr,
4th class:Power swing in 10%Pr~20%Pr,
5th class:Power swing in 20%Pr or more,
Wherein, Pr is the specified installed capacity of Wind turbines;
32) multivariate time series model D is constructedt, refer in the feature correlation for calculating multivariate time series model internal variable
It is sorted from large to small after mark MIV, and selects the corresponding variable of preceding 7 MIV as predictive variable, multivariate time series model Dt's
Data structure is:
Wherein, PRR be unit temporal power variable quantity, STD be wind speed standard deviation, Max, Min be respectively wind speed most
Big value and minimum value, Mean are wind speed average value, and subscript indicates value at the time of data;
33) using extreme learning machine ELM as classifier, using the predictive variable data after screening as extreme learning machine ELM
Input quantity the power swing of future time instance is carried out using 5 class power swing ranges as the output quantity of extreme learning machine ELM
Prediction is sorted out.
Preferably, the step 4) specifically includes following steps:
41) primary election data sample relevant to the wind power prediction moment is selected in the classification results of meteorological similitude;
42) training of the data of corresponding power swing classification as Elman neural network is selected in primary election data sample
Collection;
43) it using the network weight of BP algorithm training Elman neural network, and determines network structure, is formed and be based on sample
The tentative prediction model of similitude, and tentative prediction is carried out by the way of rolling forecast, obtain initial predicted result.
Compared with prior art, the present invention has the following advantages that:
One, the present invention proposes the power prediction side that can be applied to offshore wind farm for the prediction difficult point of offshore wind farm
Method proposes to carry out data prediction for marine wind electric field real data using PCA-SOM method, is retaining sample data as far as possible
Keynote message under the premise of carry out the preliminary classification of sample, resulting classification results have clearly classification boundaries, are conducive to
Improve model accuracy.
Two, it after preliminary classification, extracts fluctuation characteristic operating limit learning machine technology and determines prediction time wind power swing
Grade is chosen data identical with prediction time fluctuation grade on the basis of preliminary classification, is obtained further with object time
Similar historical data sample carries out being known as initial predicted model based on the power prediction of Elman network.With single Elman net
Network is compared, and initial predicted model shows the stability and superiority of precision of prediction at the time of power swing is violent.
Three, for the problem of NWP information inaccuracy, the coarse prediction of wind direction and the coarse prediction of wind speed are predicted with initial model
Input quantity of the difference as MLP, prediction error establishes marine final prediction model as output quantity, effectively improved because of
NWP information there are significant errors and caused by predicted anomaly.Show proposed model under circumstances according to numerical results
It is able to satisfy forecast demand, can be applied in actual production.
Four, when fluctuations in wind speed amplitude becomes larger, Elman model can accurately follow power swing trend to increase or subtract
It is few, but amplitude is inaccurate.And the model proposed is by Accurate Prediction wind power swing amplitude generic and with identical
The data of classification are predicted that prediction curve still can preferably follow the wave of actual power curve when wind-powered electricity generation fluctuates larger
It is dynamic, effectively increase model prediction accuracy.
Detailed description of the invention
Fig. 1 is the marine wind power prediction flow chart of the present invention.
Fig. 2 is PCA-SOM cluster result figure.
Fig. 3 is predictive variable relative importance.
Fig. 4 is that initial predicted model normalizes error.
Fig. 5 is EWSRP prediction result figure.
Fig. 6 is EWDRP prediction result figure.
Fig. 7 is that single model prediction result compares figure.
Fig. 8 is is proposed model figure compared with single Elman model prediction result.
Fig. 9 is NWP correction model correction effect figure.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment
Shanghai East Sea wind power plant is carried out for being based on meteorological similitude and the modified marine wind power prediction of NWP, such as Fig. 1
Shown, the present invention provides marine wind electric field power forecasting method, includes the following steps:
(1) due to the difference of marine monthly wind energy characteristic, in March, 2016, August and 12 months with characteristic feature are chosen
According to as research object.It, will using the first 21 days history meteorological datas of every month and wind power data as the training set of model
9 days data carry out forecast test as the test set of model afterwards, specially:
(101) the original wind energy sample chosen is expressed as S in the vector space of t momentt, data structure is as follows:
In formula, WS, WP, WD, T, W, P respectively indicate wind speed, power, wind direction, temperature, humidity and air pressure.Subscript indicates
Value at the time of data, such as WStThe wind speed forecasting provided for NWP.WSt-T1For the historical wind speed value at T1 moment before prediction time t.
(102) value of T1, T2 are sought according to the correlation of sample itself by auto-correlation function formula, as shown in formula (3).From phase
The dependence that function is used to characterize the value and moment value before at current time is closed, for stationary time series WSt, from
Correlation function is as follows:
In formula, γkAuto-correlation coefficient is represented, μ, σ respectively indicate the mean value and standard deviation of sequence;E () indicates desired value.
Finally acquiring time series T1, T2 and acquiring is 3;
(2) a kind of method for combining PCA and SOM neural network is taken when carrying out data prediction, for original
The clustering of beginning sample data.Preliminary meteorological similitude point is carried out to it while retaining sample data main feature
Class, specially:
(201) after obtaining original sample, principal component is extracted from sample using Principal Component Analysis first, forms low-dimensional
Data set.Cross-correlation matrix is established by sample data and calculates eigen vector, show that each characteristic value institute is right
Answer the contribution degree of principal component.
(202) during PCA, reduction knot of the variance contribution ratio in first three principal component as raw data set is chosen
Fruit.Its contribution rate of accumulative total is respectively 78.11%, 89.39%, 94.92%.
(203) using selected principal component as the input variable of SOM network, 4 are set by the cluster centre of SOM.Wind
Energy sample sequence cluster result is as shown in Figure 2.
(3) on the basis of dividing range to history wind power swing, when operating limit learning machine (ELM) technology will be to future
The power swing range generic at quarter predicted, specially:
(301):Power swing is fallen into 5 types by power swing amplitude.Not due to the influence factor of upper climbing and lower climbing
Together, so undulate quantity of the wind-powered electricity generation climbing grade within the unit time, is divided into following 5 by the separated classification that will climb up and down in classification
Output variable of the class as extreme learning machine:
The first kind:Power swing is below -15%Pr;
Second class:Power swing is in -15%Pr~0;
Third class:Power swing is in 0~10%Pr;
4th class:Power swing is in 10%Pr~20%Pr;
5th class:Power swing is in 20%Pr or more;
Wherein, Pr is the specified installed capacity of Wind turbines.
(302):ELM is used to carry out prediction classification as power swing of the classifier to future time instance.By above-mentioned 5 class power
Output quantity of the fluctuation range as ELM, input quantity uses data digging method, using 7 different parameters come when establishing polynary
Between series model:Average value including wind speed, standard deviation, maximum value, minimum value and power, climbing rate, NWP information.In order to
Further mining data Feature Selection PRR and the relevant preceding 5 moment history values of wind speed, data structure DtSuch as formula (7)
It is shown:
PRR represents unit time power variation in formula;STD represents the standard deviation of wind speed;Max, Min respectively represent wind speed
Maximum, minimum value;Mean represents wind speed average value;Remaining symbol is identical as sense discussed above.
(303):In order to reduce calculating cost and improve classification accuracy, the important feature that should only select initial characteristics to concentrate
As mode input amount.Optimal parameter one of of the MIV as assessment feature correlation is selected for carrying out Feature Selection to obtain
To the variable being affected to wind power swing degree.MIViIt is defined as formula (1):
The symbology related direction of MIV, absolute value represent the relative importance influenced.To the MIV value of each variable into
Row arrangement obtains the relative effect of each variable, has finally chosen 7 highest predictive variables of relative effect value.It arranges from big to small
It is classified as:Power-2, Power-4, Max-0, PRR-2, Max-2, Power-2 and Mean-1.Fig. 3 shows and arranges from big to small
30 predictive variables significance index;
(4) using tentative prediction model of the Elman neural network based on sample similarity, to the historical wind speed time
Sequence chooses the short-term wind speed rolling forecast that training sample similar with prediction time characteristic carries out wind power plant to future time instance, tool
Body is:
(401) sorted 4 class data type vector (Z1, Z2, Z3, the Z4) table of PCA and SOM Clustering Model will be passed through
Show, include the corresponding data of 5 fluctuation ranges in the data of each type, total 20 data subsets to be selected can be used as
The training data of Elman network;
(402) it first determines whether which kind of data type prediction time belongs to when carrying out wind power prediction, first assumes to belong to Z1
Class judges the wind power class of prediction time according to power swing Grade Model, is set as the 3rd class.So then from Z1 data type
In reversely extract there is the historical power of the 3rd class fluctuation grade to carry out as the trained individual Elman network of training set
Prediction, obtains initial model prediction result;
(403) using BP algorithm training network Elman network weight, determine that network structure is 7-15-1-1.Elman network
It is a typical dynamic regression network.The structure of the network is a traditional BP network, and the output of hidden layer is by holding
The delay and storage for connecing layer, are linked to the input of hidden layer certainly.Prediction result is as shown in Figure 4 as the result is shown by data clusters.
(404) rolling forecast proposed is based on rolling by hour operational process.This means that each pre-
A time available power prediction is surveyed as a result, when carrying out subsequent time prediction, by the power prediction result of last moment
As being considered as input data, while rejecting the data farthest from prediction time.The specific method is as follows:
If predicting that utilized historical series are x for the first time0(1),x0(2),…,x0(n), then the sequence of+1 rolling forecast of jth
It is classified as x0(1+j),x0(2+j),…x0(n),x0(n+1),…,x0(n+j), wherein x0(1),x0(2),…x0It (n) is history number
According to x0(n+1),…,x0(n+j), the predicted value obtained for preceding i rolling forecast.;
(5) on the basis of rudimentary model, the present invention by multilayer perceptron (MLP) to the accuracy of (NWP) information with
The inner link of model error is excavated, and NWP correction model is finally established;
The detailed process of NWP predicted error amendment based on MLP includes:
Step (501) is utilized respectively wind speed value in NWP information and wind direction predicted value be based on Elman network directly into
Its result is defined as the coarse predicted value of wind speed (WSRP) and the coarse predicted value of wind direction (WDRP) by row power prediction;
It is poor that step (502) makees WSRP, WDRP with WPP respectively, and acquired results are considered the inclined of NWP information and actual conditions
From degree, it is defined as NWP forecasting wind speed error (EWSRP) and NWP wind direction prediction error (EWDRP).Acquired results such as Fig. 5, Fig. 6
It is shown;
Step (503) by Fig. 5, Fig. 6 compared with Fig. 4, the wind speed, wind direction information and the initial predicted model that are forecast in NWP
Certain regularity is presented in error on the whole.WPR is in positively nonlinear correlation relationship with the variation of EWSRP and EWDRP, i.e.,
EWSRP and EWDRP has apparent influence to initial model prediction error.This relationship is showed using multilayer perceptron, is utilized
MLP Neural Network Self-learning function corrects NWP sequence, improves the forecast precision of NWP.
This patent selects in March, 2016, August and data in December with characteristic feature as research object;Entirety test
The data characteristics of collection is as shown in table 1:
The data characteristics in three months in 1 test set of table
23 months sea turn power prediction results of table
3 August sea wind power prediction result of table
Table sea turn power prediction in 4 December result
Fig. 7 is the result for predicting the data after screening as the input quantity of three kinds of single models.It can be with from figure
Find out that Elman model can preferably predict the trend of wind power variation, this is because its associated layers internal feedback connects
Historical data in input quantity is enhanced by Dynamic Recurrent, network generalization.But in the lower climbing of 15~18h
In this twice of event and the fluctuation of the continuous power of 25~35h, Elman model overall trend and reality are consistent but certain
The accuracy of point is not so good as BP model and RBF model.This is because model does not consider fluctuation amplitude and climbing event, cause to fluctuate
Trend is similar but climbing amplitude is different.
It is as shown in Figure 8 using Elman model and the obtained partial results of marine prediction model proposed.Become in wind speed
When changing more gentle, two kinds of models can preferably track actual wind speed curve, and prediction effect is preferable.Become in fluctuations in wind speed amplitude
When big, Elman model can accurately follow power swing trend to increase or reduce, but amplitude is inaccurate.And it is mentioned
Model out passes through Accurate Prediction wind power swing amplitude generic and is predicted with the data of the same category, in wind-powered electricity generation wave
Prediction curve still can preferably follow the fluctuation of actual power curve when moving larger, effectively increase model prediction accuracy.
The data progress NWP prediction error modified survey for finally choosing December in test set is as shown in Figure 9.NWP forecast exists
Error is larger during this section, and mean absolute error has reached 3.13m/s.The initial predicted model proposed respectively using 3.3 sections
It is predicted respectively with what is proposed by the modified final mask of NWP correction module.Predict that error is filled with wind field in Fig. 9
Machine capacity does normalized, it can be seen that the predicted anomaly point of initial predicted model is more, has the error of certain points very
To having reached 79%.This is because December, entirety wind speed was larger, and when middle high wind speed, according to power of fan curve NWP
The error of wind speed forecasting and wind direction forecast very little brings biggish power prediction error.In addition, wind direction forecast has in high wind speed
Small difference can similarly make to predict that error increases.In Fig. 9, the prediction result finally predicted is in abnormal point, catastrophe point
Correction effect is obvious, is also improved in the wind power relatively gentle stage.It should be pointed out that certain prediction accuracies compared with
Error will increase instead after high point is corrected by correction model, but the range increased is within the scope of acceptable.
Claims (5)
1. a kind of marine wind electric field power forecasting method for considering meteorological similitude and power swing, which is characterized in that including with
Lower step:
1) the wind energy data in marine wind electric field certain time are obtained, and as original wind energy sample;
2) Principal Component Analysis and self organizing neural network is combined to carry out the cluster of meteorological similitude to original wind energy sample data
Analysis, obtains meteorological similitude classification results;
3) classify to power swing, and obtain the generic of power swing range using extreme learning machine ELM, obtain function
The classification results of rate fluctuation;
4) using tentative prediction model of the Elman neural network based on sample similarity, classified according to meteorological similitude
As a result training sample relevant to prediction time being chosen with the classification results of power swing, the short of wind power plant is carried out to future time instance
Phase wind speed rolling forecast, and obtain initial predicted result;
5) NWP correction model is established using multilayer perceptron MLP to be modified initial predicted result, obtain revised final
As a result.
2. a kind of marine wind electric field power prediction side for considering meteorological similitude and power swing according to claim 1
Method, which is characterized in that in the step 1), original wind energy sample is expressed as S in the vector space of t momentt, data structure
For:
Wherein, WS, WP, WD, T, W, P respectively indicate wind speed, power, wind direction, temperature, humidity and barometric information, and subscript indicates
Value at the time of data, WStFor the wind speed forecasting data of prediction time t, WSt-T1For the historical wind speed at T1 moment before prediction time t
Value.
3. a kind of marine wind electric field power prediction side for considering meteorological similitude and power swing according to claim 1
Method, which is characterized in that the step 2) specifically includes following steps:
21) principal component is extracted by Principal Component Analysis, and obtains the contribution rate of corresponding principal component by cross-correlation matrix, and select
Taking cumulative proportion in ANOVA is more than 90% corresponding principal component;
22) using the principal component after screening as the input variable of self organizing neural network, cluster centre quantity is set, is carried out similar
Property cluster to obtain preliminary meteorological similitude classification.
4. a kind of marine wind electric field power prediction side for considering meteorological similitude and power swing according to claim 2
Method, which is characterized in that the step 3) specifically includes following steps:
31) power swing is fallen into 5 types by power swing amplitude, specially:
The first kind:Power swing in -15%Pr hereinafter,
Second class:Power swing in -15%Pr~0,
Third class:Power swing in 0~10%Pr,
4th class:Power swing in 10%Pr~20%Pr,
5th class:Power swing in 20%Pr or more,
Wherein, Pr is the specified installed capacity of Wind turbines;
32) multivariate time series model D is constructedt, in the feature correlation index MIV for calculating multivariate time series model internal variable
After sort from large to small, and select the corresponding variable of preceding 7 MIV as predictive variable, multivariate time series model DtData
Structure is:
Wherein, PRR is unit temporal power variable quantity, and STD is the standard deviation of wind speed, and Max, Min are respectively the maximum value of wind speed
And minimum value, Mean are wind speed average value, subscript indicates value at the time of data;
33) using extreme learning machine ELM as classifier, using the predictive variable data after screening as the defeated of extreme learning machine ELM
Enter amount, using 5 class power swing ranges as the output quantity of extreme learning machine ELM, the power swing of future time instance is predicted
Sort out.
5. a kind of marine wind electric field power prediction side for considering meteorological similitude and power swing according to claim 1
Method, which is characterized in that the step 4) specifically includes following steps:
41) primary election data sample relevant to the wind power prediction moment is selected in the classification results of meteorological similitude;
42) training set of the data of corresponding power swing classification as Elman neural network is selected in primary election data sample;
43) it using the network weight of BP algorithm training Elman neural network, and determines network structure, is formed similar based on sample
The tentative prediction model of property, and tentative prediction is carried out by the way of rolling forecast, obtain initial predicted result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810696196.0A CN108898251B (en) | 2018-06-29 | 2018-06-29 | Offshore wind farm power prediction method considering meteorological similarity and power fluctuation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810696196.0A CN108898251B (en) | 2018-06-29 | 2018-06-29 | Offshore wind farm power prediction method considering meteorological similarity and power fluctuation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108898251A true CN108898251A (en) | 2018-11-27 |
CN108898251B CN108898251B (en) | 2021-10-08 |
Family
ID=64347021
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810696196.0A Active CN108898251B (en) | 2018-06-29 | 2018-06-29 | Offshore wind farm power prediction method considering meteorological similarity and power fluctuation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108898251B (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109784473A (en) * | 2018-12-13 | 2019-05-21 | 天津大学 | A kind of short-term wind power prediction method based on Dual Clocking feature learning |
CN110276472A (en) * | 2019-05-21 | 2019-09-24 | 南方电网调峰调频发电有限公司 | A kind of offshore wind farm power ultra-short term prediction method based on LSTM deep learning network |
CN110942170A (en) * | 2019-08-29 | 2020-03-31 | 华北电力大学(保定) | Short-term wind speed prediction method and system based on information processing |
CN111008725A (en) * | 2019-10-23 | 2020-04-14 | 中国农业大学 | Meteorological factor fluctuation feature extraction method for short-term wind power prediction |
CN111030106A (en) * | 2019-12-26 | 2020-04-17 | 福州大学 | Wind power fluctuation quality assessment method based on waveform similarity theory |
CN111626473A (en) * | 2020-04-17 | 2020-09-04 | 国网上海市电力公司 | Two-stage photovoltaic power prediction method considering error correction |
CN112149879A (en) * | 2020-09-02 | 2020-12-29 | 上海电力大学 | New energy medium-and-long-term electric quantity prediction method considering macroscopic volatility classification |
CN112216181A (en) * | 2020-10-30 | 2021-01-12 | 哈尔滨飞机工业集团有限责任公司 | System and method capable of realizing single-shot training of double-shot helicopter |
CN112288164A (en) * | 2020-10-29 | 2021-01-29 | 四川大学 | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast |
CN112418346A (en) * | 2020-12-07 | 2021-02-26 | 天津大学 | Numerical weather forecast total radiation system error classification calculation method |
CN113449847A (en) * | 2021-06-17 | 2021-09-28 | 江苏方天电力技术有限公司 | Offshore wind power rolling prediction method considering second-level time series wind speed change |
CN114297186A (en) * | 2021-12-30 | 2022-04-08 | 广西电网有限责任公司 | Power consumption data preprocessing method and system based on deviation coefficient |
CN114465231A (en) * | 2022-02-14 | 2022-05-10 | 南京邮电大学盐城大数据研究院有限公司 | Power access method and system for offshore wind power generation |
CN117937638A (en) * | 2024-03-12 | 2024-04-26 | 山西城市动力新能源有限公司 | Wind power plant cluster power generation control method and device, electronic equipment and storage medium |
CN117937638B (en) * | 2024-03-12 | 2024-07-09 | 山西城市动力新能源有限公司 | Wind power plant cluster power generation control method and device, electronic equipment and storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011124226A1 (en) * | 2010-04-08 | 2011-10-13 | Vestas Wind Systems A/S | Method and system for forecasting wind energy |
CN105117788A (en) * | 2015-07-22 | 2015-12-02 | 河南行知专利服务有限公司 | Wind power generation power prediction method |
CN105894106A (en) * | 2015-01-05 | 2016-08-24 | 国家电网公司 | Integral coupling method of ocean model and meteorological model |
CN105930900A (en) * | 2016-05-09 | 2016-09-07 | 华北电力大学 | Method and system for predicting hybrid wind power generation |
EP3113312A1 (en) * | 2015-07-01 | 2017-01-04 | General Electric Company | Predictive control for energy storage on a renewable energy system |
CN107067099A (en) * | 2017-01-25 | 2017-08-18 | 清华大学 | Wind power probability forecasting method and device |
CN107609697A (en) * | 2017-09-06 | 2018-01-19 | 南京邮电大学 | A kind of Wind power forecasting method |
CN107944622A (en) * | 2017-11-21 | 2018-04-20 | 华北电力大学 | Wind power forecasting method based on continuous time cluster |
CN108133279A (en) * | 2017-08-29 | 2018-06-08 | 甘肃省电力公司风电技术中心 | Wind power probability forecasting method, storage medium and equipment |
CN108205717A (en) * | 2017-12-30 | 2018-06-26 | 国网江苏省电力公司无锡供电公司 | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology |
-
2018
- 2018-06-29 CN CN201810696196.0A patent/CN108898251B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011124226A1 (en) * | 2010-04-08 | 2011-10-13 | Vestas Wind Systems A/S | Method and system for forecasting wind energy |
CN105894106A (en) * | 2015-01-05 | 2016-08-24 | 国家电网公司 | Integral coupling method of ocean model and meteorological model |
EP3113312A1 (en) * | 2015-07-01 | 2017-01-04 | General Electric Company | Predictive control for energy storage on a renewable energy system |
CN105117788A (en) * | 2015-07-22 | 2015-12-02 | 河南行知专利服务有限公司 | Wind power generation power prediction method |
CN105930900A (en) * | 2016-05-09 | 2016-09-07 | 华北电力大学 | Method and system for predicting hybrid wind power generation |
CN107067099A (en) * | 2017-01-25 | 2017-08-18 | 清华大学 | Wind power probability forecasting method and device |
CN108133279A (en) * | 2017-08-29 | 2018-06-08 | 甘肃省电力公司风电技术中心 | Wind power probability forecasting method, storage medium and equipment |
CN107609697A (en) * | 2017-09-06 | 2018-01-19 | 南京邮电大学 | A kind of Wind power forecasting method |
CN107944622A (en) * | 2017-11-21 | 2018-04-20 | 华北电力大学 | Wind power forecasting method based on continuous time cluster |
CN108205717A (en) * | 2017-12-30 | 2018-06-26 | 国网江苏省电力公司无锡供电公司 | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology |
Non-Patent Citations (3)
Title |
---|
AOIFE M. FOLEY等: "Current methods and advances in forecasting of wind power generation", 《RENEWABLE ENERGY》 * |
彭小圣等: "风电集群短期及超短期功率预测精度改进方法综述", 《中国电机工程学报》 * |
迟永宁等: "大规模海上风电输电与并网关键技术研究综述", 《中国电机工程学报》 * |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109784473A (en) * | 2018-12-13 | 2019-05-21 | 天津大学 | A kind of short-term wind power prediction method based on Dual Clocking feature learning |
CN110276472A (en) * | 2019-05-21 | 2019-09-24 | 南方电网调峰调频发电有限公司 | A kind of offshore wind farm power ultra-short term prediction method based on LSTM deep learning network |
CN110942170A (en) * | 2019-08-29 | 2020-03-31 | 华北电力大学(保定) | Short-term wind speed prediction method and system based on information processing |
CN111008725A (en) * | 2019-10-23 | 2020-04-14 | 中国农业大学 | Meteorological factor fluctuation feature extraction method for short-term wind power prediction |
CN111008725B (en) * | 2019-10-23 | 2022-06-24 | 中国农业大学 | Meteorological factor fluctuation feature extraction method for short-term wind power prediction |
CN111030106B (en) * | 2019-12-26 | 2021-04-27 | 福州大学 | Wind power fluctuation quality assessment method based on waveform similarity theory |
CN111030106A (en) * | 2019-12-26 | 2020-04-17 | 福州大学 | Wind power fluctuation quality assessment method based on waveform similarity theory |
CN111626473A (en) * | 2020-04-17 | 2020-09-04 | 国网上海市电力公司 | Two-stage photovoltaic power prediction method considering error correction |
CN112149879A (en) * | 2020-09-02 | 2020-12-29 | 上海电力大学 | New energy medium-and-long-term electric quantity prediction method considering macroscopic volatility classification |
CN112149879B (en) * | 2020-09-02 | 2022-10-18 | 上海电力大学 | New energy medium-and-long-term electric quantity prediction method considering macroscopic volatility classification |
CN112288164B (en) * | 2020-10-29 | 2023-04-07 | 四川大学 | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast |
CN112288164A (en) * | 2020-10-29 | 2021-01-29 | 四川大学 | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast |
CN112216181A (en) * | 2020-10-30 | 2021-01-12 | 哈尔滨飞机工业集团有限责任公司 | System and method capable of realizing single-shot training of double-shot helicopter |
CN112418346A (en) * | 2020-12-07 | 2021-02-26 | 天津大学 | Numerical weather forecast total radiation system error classification calculation method |
CN112418346B (en) * | 2020-12-07 | 2022-10-25 | 天津大学 | Numerical weather forecast total radiation system error classification calculation method |
CN113449847A (en) * | 2021-06-17 | 2021-09-28 | 江苏方天电力技术有限公司 | Offshore wind power rolling prediction method considering second-level time series wind speed change |
CN114297186A (en) * | 2021-12-30 | 2022-04-08 | 广西电网有限责任公司 | Power consumption data preprocessing method and system based on deviation coefficient |
CN114297186B (en) * | 2021-12-30 | 2024-04-26 | 广西电网有限责任公司 | Power consumption data preprocessing method and system based on deviation coefficient |
CN114465231A (en) * | 2022-02-14 | 2022-05-10 | 南京邮电大学盐城大数据研究院有限公司 | Power access method and system for offshore wind power generation |
CN114465231B (en) * | 2022-02-14 | 2023-06-13 | 南京邮电大学盐城大数据研究院有限公司 | Power access method and system for offshore wind power generation |
CN117937638A (en) * | 2024-03-12 | 2024-04-26 | 山西城市动力新能源有限公司 | Wind power plant cluster power generation control method and device, electronic equipment and storage medium |
CN117937638B (en) * | 2024-03-12 | 2024-07-09 | 山西城市动力新能源有限公司 | Wind power plant cluster power generation control method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN108898251B (en) | 2021-10-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898251A (en) | Consider the marine wind electric field power forecasting method of meteorological similitude and power swing | |
Lipu et al. | Artificial intelligence based hybrid forecasting approaches for wind power generation: Progress, challenges and prospects | |
CN111753893A (en) | Wind turbine generator power cluster prediction method based on clustering and deep learning | |
Kabir et al. | Predicting thunderstorm-induced power outages to support utility restoration | |
CN108539738B (en) | Short-term load prediction method based on gradient lifting decision tree | |
CN103268366B (en) | A kind of combination wind power forecasting method suitable for distributing wind power plant | |
US20230023931A1 (en) | Hydraulic turbine cavitation acoustic signal identification method based on big data machine learning | |
CN110807550B (en) | Distribution transformer overload recognition and early warning method based on neural network and terminal equipment | |
CN111488896B (en) | Distribution line time-varying fault probability calculation method based on multi-source data mining | |
CN113379142B (en) | Short-term wind power prediction method based on wind speed correction and fusion model | |
CN114399081A (en) | Photovoltaic power generation power prediction method based on weather classification | |
CN113610285A (en) | Power prediction method for distributed wind power | |
CN108038518A (en) | A kind of photovoltaic generation power based on meteorological data determines method and system | |
CN116663393A (en) | Random forest-based power distribution network continuous high-temperature fault risk level prediction method | |
CN116187540B (en) | Wind power station ultra-short-term power prediction method based on space-time deviation correction | |
CN114997475B (en) | Kmeans-based fusion model photovoltaic power generation short-term prediction method | |
Xiong et al. | An ultra-short-term wind speed correction method based on the fluctuation characteristics of wind speed | |
CN115496264A (en) | Method for predicting generated power of wind turbine generator | |
アンナススワルディ et al. | Neuro-fuzzy approaches for modeling the wet season tropical rainfall | |
CN113723670A (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
Aweda et al. | Investigation of wind speed to generate energy using machine learning algorithms approach over selected Nigerian Stations | |
Ren et al. | A study of wind speed prediction based on particle swarm algorithm to optimize the parameters of sparse least squares support vector | |
CN117633456B (en) | Marine wind power weather event identification method and device based on self-adaptive focus loss | |
Zhu et al. | Resilience-Oriented Extreme Weather Conditional Renewable Scenario Generation Based on Diffusion Models and Few-shot Learning | |
CN117290684B (en) | Transformer-based high-temperature drought weather early warning method and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |