CN107358006A - A kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis - Google Patents
A kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis Download PDFInfo
- Publication number
- CN107358006A CN107358006A CN201710613264.8A CN201710613264A CN107358006A CN 107358006 A CN107358006 A CN 107358006A CN 201710613264 A CN201710613264 A CN 201710613264A CN 107358006 A CN107358006 A CN 107358006A
- Authority
- CN
- China
- Prior art keywords
- wind speed
- disturbance
- lorenz
- principal component
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Computer Hardware Design (AREA)
- Marketing (AREA)
- Geometry (AREA)
- Game Theory and Decision Science (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
Abstract
The invention discloses a kind of Lorenz based on principal component analysis to disturb wind speed forecasting method, belongs to technical field of power generation.The present invention is to carry out principal component analysis according to initial data, least square method supporting vector machine (LS SVM) model is recycled to carry out forecasting wind speed to principal component and air speed data, finally it is modified using the initial predicted value of Lorenz atmospheric perturbation sequence pair wind speed, to improve the prediction level of wind speed.Simulation result is shown, the disturbing influence of Atmosphere System is taken into full account during forecasting wind speed, significantly improves the precision of prediction of wind speed.The present invention compensate for ignorance effect of the forecasting wind speed field to Atmosphere System, contribute to the stability of wind-electricity integration and large-scale developing and utilizing for wind-resources.
Description
Technical field
The present invention relates to a kind of method for being capable of Accurate Prediction wind farm wind velocity, belong to technical field of power generation.
Background technology
The world today, a series of problems, such as supply anxiety of fossil fuel brings environment, ecology and Global climate change,
Each national capital conscious action is got up, and actively cracks quagmire, accelerates energy transition, Renewable Energy Development.A kind of environmental protection of wind energy conduct,
Reproducible clean energy resource, very important status is occupied in low-carbon energy technology, receive the extensive concern in the whole world.Together
When, as share of the wind-power electricity generation in power generation is increasing, grid operator is in the power balance, quality of power supply, grid-connected
Stability, load scheduling plan etc. Challenge.Therefore, wind-power electricity generation it is grid-connected during, it is necessary to consideration can
The wind power forecasting method leaned on.
Domestic and international researcher proposes much research methods on wind power prediction.Wind power prediction is pressed at present
It is divided into long-term forecast, medium-term forecast, short-term forecast and ultra-short term prediction according to the length of predetermined period;According to the mathematical modeling of foundation
Physical model, statistical model, artificial intelligence technology and built-up pattern can be divided into.In view of the diversity of Forecasting Methodology, the country is outgoing
Many mature and stable forecast systems are showed.It is more early that foreign countries are engaged in Wind power forecasting research work, most representational
Forecast work includes Prediktor, WPPT, Previento, eWind etc..The wind power forecast system of domestic-developed is main
There are WINPOP systems, WPPS systems, WPFS systems, T213L31 systems etc..
Wind power forecasting method above is the linguistic term based on algorithm model and forecast system mostly, and is ignored
Studied for the disturbing influence of big pneumatic power system, thus prediction effect is less desirable.
The content of the invention
The technical problems to be solved by the invention are to provide a kind of Lorenz disturbance forecasting wind speeds based on principal component analysis
Method, to reach the purpose for improving forecasting wind speed.
In order to solve the above technical problems, the technical solution used in the present invention is:It is a kind of based on principal component analysis
Lorenz disturbs wind speed forecasting method, it is characterised in that:
(1) 1 group of initial data is gathered every one section of duration t' to wind power plant to be measured, n groups original is collected at moment time t-1
Beginning data, the initial data include air speed data (v)=(vt-1-(n-1)t',vt-1-(n-2)t',......,vt-1) and to wind speed
Influence factor initial dataX1,X2,......,XmThe shadow to wind speed is represented respectively
Set of factors is rung, m represents the influence factor number to wind speed;
(2) the influence factor initial data (X) to wind speed in the n group initial data for being collected the t-1 moment is carried out
Principal component analysis, number of principal components evidence is determined according to contribution rate of accumulative total;
(3) number of principal components is combined according to the air speed data (v) in the n group initial data with the t-1 moment is collected
As the input of least square method supporting vector machine (LS-SVM) model, by the wind speed v of ttOutput as model is to data
It is trained, finally obtains the tentative prediction sequence v'(t of wind speed);
(4) initial condition and parameter are given, solves Lorenz equations, obtains the atmospheric perturbation sequence of three-dimensional, and to disturbance
Sequence does standardization, obtains the standardization disturbance sequence of countless magnitudes and dimension;
(5) second order Minkowski (Minkowski) distance between normalized data and initial value, is obtained
Lorenz synthesis disturbance streams;
(6) tentative prediction sequence v'(t of the disturbance stream to wind speed is integrated using Lorenz) disturbance amendment is carried out, and to realize
The minimum target of mean absolute error between wind speed actual value and predicted value obtains the Optimal Disturbance intensity and most of disturbance amendment
Excellent coefficient of disturbance;
(7) the Optimal Disturbance intensity and Optimal Disturbance coefficient obtained using Lorenz synthesis disturbance streams is to the preliminary pre- of wind speed
Sequencing row carry out disturbance amendment, obtain the disturbance Orders Corrected v " (t) of wind speed.
Further technical scheme is, the influence factor to wind speed include wind direction D, temperature T, air pressure P, specific volume a,
Than wet H and roughness of ground surface R;
Further technical scheme is that the process of the principal component analysis is as follows:
A, with wind direction D sine value, wind direction D cosine value, temperature T, air pressure P, specific volume a, than wet H and roughness of ground surface R
For the influence factor to wind speed;The data of collection are built to the initial data of principal component analysis, it is as follows:
X1,X2,......,XmWind direction D sine value, wind direction D is represented respectively
Cosine value, temperature T, air pressure P, specific volume a, the set of factors than wet H and roughness of ground surface R, m=7 here;
B, to n × 7 initial data (X) standardization of 7 set of factors, to eliminate the difference between variable on the order of magnitude
Influenceed caused by different, the average for making each variable is 0, variance 1, obtains normalized matrix Y={ ynm};
Wherein, xmAnd SmX is represented respectivelymAverage and variance.
C, correlation matrix A is drawn according to normalized matrix;
D, correlation matrix A characteristic value and characteristic vector are calculated, and calculates eigenvalue contribution rate and contribution rate of accumulative total;
Wherein, λmRepresent A characteristic value.
E, principal component Z=[Z are asked according to contribution rate of accumulative total1,Z2,...,Zm]
Zm=Yam
Wherein, amRepresent λmCharacteristic vector;m≤7.
Further technical scheme is that the step (3) is by number of principal components according to the n with the t-1 moment is collected
Air speed data (v) in group initial data is combined as the input matrix M of least square method supporting vector machine model, by t
Wind speed vtOutput matrix N as model is trained to data, finally obtains the tentative prediction sequence v'(t of wind speed);
Wherein,
In formula, dmnRepresent the data that principal component obtains;
Further technical scheme is that the Lorenz equations are as follows:
Wherein, x in formula, y, z are state variables, represent intensity, the level side of convective motion fluid of convective motion respectively
To the temperature difference and vertical temperature-difference to without convection current when departure degree.σ, r, b are nondimensional arithmetic numbers, represent Prandtl respectively
Number, Rayleigh number and the amount related to climatic province size.For classics Lorenz systems parameter value σ=10,Give
Determine initial value h=(1.1,1,1), make r=23, r=27, we can obtain two groups of different numerical solutions, and then obtain two groups of differences
Lorenz attractors.Because the numerical solution of Lorenz equations is one group of three-dimensional disturbance sequence, thus first non trivial solution is carried out
Standardization, obtain the standardization disturbance sequence of countless magnitudes and dimensionSpecific standardsization handle formula:
X in formulan,yn,znRepresent the numerical solution of Lorenz equations, xmin,ymin,zminAnd xmax,ymax,zmaxX is represented respectively,
Y, z minimum value and maximum.
Further technical scheme is that the distance of second order Minkowski (Minkowski) described in the step (5) is as follows:
Second order Minkowski Distance between normalized data and initial value (0,0,0):
Wherein, CnRepresentC0Represent C0(x0,y0,z0)=C0(0,0,0)。
Further technical scheme is, using mean absolute error MAE and root-mean-square error RMSE come quantitative evaluation
The quality of model, error amount is smaller, represents that the predictive ability of model is better.
Its formula is
Wherein, y (t) and v (t) represents the actual value and predicted value of wind speed respectively, and k is forecast sample number.
It is using beneficial effect caused by above-mentioned technical proposal:Description air motion is introduced during forecasting wind speed
The Lorenz systems of state carry out disturbance amendment to forecasting wind speed result, and carry out emulation mould using the real data of wind power plant
Intend.Test result indicates that significantly improving the prediction level of wind power plant by the revised model of Lorenz disturbances, contribute to wind
The large-scale development of energy resource and utilization.
Brief description of the drawings
The present invention is further detailed explanation with reference to the accompanying drawings and detailed description.
Fig. 1 (a) and (b) are respectively wind speed profile situation and its wind rose map of the wind power plant every ten minutes;
Fig. 2 (a) and (b) are distributed as Lorenz attractor morphologies during Rayleigh number;
Fig. 3 is one group of Lorenz synthesis disturbance streams that Lorenz systems generate under Euclidean distance;
Lorenz strength of turbulences when Fig. 4 is, abscissa represent the data amount check of disturbance, i.e. forecast sample data, indulge and sit
Mark represents the size of strength of turbulence;
Fig. 5 is the forecasting wind speed curve of each model;
Fig. 6 is the flow chart of the present invention.
Embodiment
With reference to the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Ground describes, it is clear that described embodiment is only the part of the embodiment of the present invention, rather than whole embodiments.Based on this
Embodiment in invention, the every other reality that those of ordinary skill in the art are obtained under the premise of creative work is not made
Example is applied, belongs to the scope of protection of the invention.
Many details are elaborated in the following description to facilitate a thorough understanding of the present invention, still the present invention can be with
It is different from other manner described here using other to implement, those skilled in the art can be without prejudice to intension of the present invention
In the case of do similar popularization, therefore the present invention is not limited by following public specific embodiment.
The invention provides a kind of Lorenz based on principal component analysis to disturb wind speed forecasting method, and one kind is based on principal component
The Lorenz disturbance wind speed forecasting methods of analysis, it is characterised in that:
(1) 1 group of initial data is gathered every one section of duration t' to wind power plant to be measured, n groups original is collected at moment time t-1
Beginning data, the initial data include air speed data (v)=(vt-1-(n-1)t',vt-1-(n-2)t',......,vt-1) and to wind speed
Influence factor initial dataX1,X2,......,XmThe shadow to wind speed is represented respectively
Set of factors is rung, m represents the influence factor number to wind speed;
Wind direction D, temperature T, air pressure P, specific volume a, than wet H and earth's surface more coarse can be selected the influence factor of wind speed herein
Spend R.
(2) the influence factor initial data (X) to wind speed in the n group initial data for being collected the t-1 moment is made
For principal component analysis, number of principal components evidence is determined according to contribution rate of accumulative total;
The process of the principal component analysis is as follows:
A, with wind direction D sine value, wind direction D cosine value, temperature T, air pressure P, specific volume a, than wet H and roughness of ground surface R
For the influence factor to wind speed;The data of collection are built to the initial data of principal component analysis, it is as follows:
X1,X2,......,XmWind direction D sine value, wind direction D is represented respectively
Cosine value, temperature T, air pressure P, specific volume a, the set of factors than wet H and roughness of ground surface R, m=7 here;
B, to n × 7 initial data (X) standardization of 7 set of factors, to eliminate the difference between variable on the order of magnitude
Influenceed caused by different, the average for making each variable is 0, variance 1, obtains normalized matrix Y={ ynm};
Wherein,And SmX is represented respectivelymAverage and variance.
C, correlation matrix A is drawn according to normalized matrix;
D, correlation matrix A characteristic value and characteristic vector are calculated, and calculates eigenvalue contribution rate and contribution rate of accumulative total;
Wherein, λmRepresent A characteristic value.
E, principal component Z=[Z are asked according to contribution rate of accumulative total1,Z2,...,Zm]
Zm=Yam
Wherein, amRepresent λmCharacteristic vector;m≤7.
(3) number of principal components is combined according to the air speed data (v) in the n group initial data with the t-1 moment is collected
As the input of least square method supporting vector machine model, by the wind speed v of ttOutput as model is trained to data,
Finally obtain the tentative prediction sequence v'(t of wind speed);
Number of principal components is combined as according to the air speed data (v) in the n group initial data with the t-1 moment is collected
The input matrix M of least square method supporting vector machine model, by the wind speed v of ttOutput matrix N as model enters to data
Row training, finally obtain the tentative prediction sequence v'(t of wind speed);
Wherein,
In formula, dmnRepresent the data that principal component obtains;
(4) initial condition and parameter are given, solves Lorenz equations, obtains the atmospheric perturbation sequence of three-dimensional, and to disturbance
Sequence does standardization, obtains the standardization disturbance sequence of countless magnitudes and dimension;
The Lorenz equations are as follows:
Wherein, x in formula, y, z are state variables, represent intensity, the level side of convective motion fluid of convective motion respectively
To the temperature difference and vertical temperature-difference to without convection current when departure degree.σ, r, b are nondimensional arithmetic numbers, represent Prandtl respectively
Number, Rayleigh number and the amount related to climatic province size.For classics Lorenz systems parameter value σ=10,Give
Determine initial value h=(1.1,1,1), make r=23, r=27, we can obtain two groups of different numerical solutions, and then obtain two groups of differences
Lorenz attractors.Because the numerical solution of Lorenz equations is one group of three-dimensional disturbance sequence, thus first non trivial solution is entered
Row standardization, obtain the standardization disturbance sequence of countless magnitudes and dimensionSpecific standardsization handle formula:
X in formulan,yn,znRepresent the numerical solution of Lorenz equations, xmin,ymin,zminAnd xmax,ymax,zmaxX is represented respectively,
Y, z minimum value and maximum.
(5) the second order Minkowski Distance between normalized data and initial value, Lorenz synthesis disturbances are obtained
Stream;
Second order Minkowski (Minkowski) distance is as follows:
Second order Minkowski Distance between normalized data and initial value (0,0,0):
Wherein, CnRepresentC0Represent C0(x0,y0,z0)=C0(0,0,0)。
(6) the Lorenz synthesis disturbance streams that (5) obtain are utilized to the least square method supporting vector machine based on principal component analysis
Obtained wind speed tentative prediction sequence is modified, and disturbance correction formula is:
v″(t1,t2,...,tk)=v ' (t1,t2,...,tk)+ld(t1,t2,...,tk)
Wherein, v " (t1,t2,...,tk) be wind speed disturbance Orders Corrected;v′(t1,t2,...,tk) it is the preliminary of wind speed
Forecasting sequence.L is coefficient of disturbance, if l is positive number, shows that wind speed tentative prediction sequence needs to strengthen disturbance;If l is negative,
Then show that wind speed tentative prediction sequence needs to weaken disturbance.d(t1,t2,...,tk) it is strength of turbulence, represent Lorenz synthesis and disturb
The partial sequence of dynamic stream.K is the sample number of prediction.
(7) tentative prediction sequence v'(t of the disturbance stream to wind speed is integrated using Lorenz) disturbance amendment is carried out, and to realize
The minimum target of mean absolute error between wind speed actual value and predicted value obtains the Optimal Disturbance intensity and most of disturbance amendment
Excellent coefficient of disturbance;
(8) the Optimal Disturbance intensity and Optimal Disturbance coefficient obtained using Lorenz synthesis disturbance streams is to the preliminary pre- of wind speed
Sequencing row carry out disturbance amendment, obtain the disturbance Orders Corrected v " (t) of wind speed.
Based on this Forecasting Methodology, inventor additionally provide how the method for evaluation and foreca effect;It uses average absolute to miss
Poor MAE and root-mean-square error RMSE carrys out the quality of quantitative evaluation model, and error amount is smaller, represents that the predictive ability of model is got over
It is good.
Its formula is
Wherein, y (t) and v (t) represents the actual value and predicted value of wind speed respectively, and k is forecast sample number.
LS-SVM models, the LS-SVM models (P-LSSVM) of principal component analysis and Lorenz are disturbed using error criterion
The prediction level of the LS-SVM models (L-P-LSSVM) of the principal component analysis of amendment carries out quantitative analysis.
Embodiment
In collection wind power plant every the initial data of observation in 10 minutes, including wind speed, wind direction data and cubic spline is used
Temperature that interpolation obtains, air pressure, specific volume, than data such as wet and roughness of ground surface;Carry out relevant treatment.
Fig. 1 (a) and (b) are respectively wind speed profile situation and its wind rose map of the wind power plant every ten minutes, can be seen
There is typical stochastic volatility to the wind speed in figure (a);It is most in section (6,8) wind speed to scheme the wind speed and direction rose figure of (b),
It is most long in southeastern direction colour band, represent the wind direction frequency highest.When Fig. 2 (a) and (b) are distributed as Rayleigh number r=23, r=27
Lorenz attractor morphologies, it can be seen that for same initial value, different Rayleigh numbers can obtain different Lorenz and attract
Sub- form.Fig. 3 is the Lorenz synthesis disturbance streams of Lorenz systems generations different under Euclidean distance, and Lorenz integrates disturbance
Stream still has very strong randomness.Fig. 4 is r=23, Lorenz coefficients of disturbance and strength of turbulence during r=27, abscissa table
Show the data amount check of disturbance, i.e. forecast sample data, ordinate represents the size of strength of turbulence, and the l in legend represents disturbance system
Number.
Wind speed is predicted according to the modeling procedure of model above, obtains the forecasting wind speed curve of each model, such as Fig. 5
It is shown.As we can see from the figure compared to least square method supporting vector machine model, the least square by principal component analysis is supported
The forecasting wind speed curve of vector machine model and the actual trend of wind speed are consistent;The relatively low master of overall forecasting wind speed sequence simultaneously
The forecasting wind speed model of constituent analysis, which needs to apply, just to be disturbed so that it more meets the actual value of wind speed.Pink colour and green solid lines table
Show the forecasting wind speed result of Lorenz disturbance amendments, the Lorenz equations of different Rayleigh numbers disturb revised model closer to reality
Border wind series, there is more preferable prediction result.The optimal L orenz disturbance sequences obtained under this two kinds of Lorenz equation state of explanation
Row all reduce the randomness of wind series, have reached the purpose for improving forecasting wind speed result.
Quantitative analysis is carried out to each model using MAE and RMSE error criterions, specific precision of prediction is as shown in table 1.
In the quality of model, error criterion data and the forecasting wind speed figure of each model maintain highly consistent.Based on principal component most
The error criterion data that a young waiter in a wineshop or an inn multiplies supporting vector machine model are slightly less than single LS-SVM models, while each mistake of the two models
Poor index data are far longer than through the revised forecasting wind speed models of Lorenz.This does not only illustrate the advantage of principal component analysis,
It also illustrate that good improvement effect of the Lorenz equations to wind speed tentative prediction sequence.The different Lorenz disturbance sequences of Rayleigh number
The randomness during forecasting wind speed is have modified, significantly improves the prediction level of wind speed.
The error analysis of 1 each model of table
Claims (7)
- A kind of 1. Lorenz disturbance wind speed forecasting methods based on principal component analysis, it is characterised in that:(1) 1 group of initial data is gathered every one section of duration t' to wind power plant to be measured, n group original numbers is collected at moment time t-1 According to the initial data includes air speed data (v)=(vt-1-(n-1)t',vt-1-(n-2)t',......,vt-1) and shadow to wind speed The factor of sound initial dataX1,X2,......,XmRepresent respectively influence to wind speed because Element collection, m represent the influence factor number to wind speed;(2) the influence factor initial data (X) to wind speed in the n group initial data for being collected the t-1 moment is as master Constituent analysis, number of principal components evidence is determined according to contribution rate of accumulative total;(3) number of principal components is combined as according to the air speed data (v) in the n group initial data with the t-1 moment is collected The input of least square method supporting vector machine model, by the wind speed v of ttOutput as model is trained to data, finally Obtain the tentative prediction sequence v'(t of wind speed);(4) initial condition and parameter are given, solves Lorenz equations, obtains the atmospheric perturbation sequence of three-dimensional, and to disturbing sequence Standardization is done, obtains the standardization disturbance sequence of countless magnitudes and dimension;(5) the second order Minkowski Distance between normalized data and initial value, Lorenz synthesis disturbance streams are obtained;(6) tentative prediction sequence v'(t of the disturbance stream to wind speed is integrated using Lorenz) disturbance amendment is carried out, and to realize wind speed The minimum target of mean absolute error between actual value and predicted value obtains the Optimal Disturbance intensity of disturbance amendment and optimal disturbed Dynamic coefficient;(7) tentative prediction sequence of the Optimal Disturbance intensity and Optimal Disturbance coefficient obtained using Lorenz synthesis disturbance streams to wind speed Row carry out disturbance amendment, obtain the disturbance Orders Corrected v " (t) of wind speed.
- 2. according to claim 1 a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis, its feature exists In:The influence factor to wind speed include wind direction D, temperature T, air pressure P, specific volume a, than wet H and roughness of ground surface R.
- 3. according to claim 1 a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis, its feature exists In:The process of the principal component analysis is as follows:A, using wind direction D sine value, wind direction D cosine value, temperature T, air pressure P, specific volume a, than wet H and roughness of ground surface R as pair The influence factor of wind speed;The data of collection are built to the initial data of principal component analysis, it is as follows:X1,X2,......,XmWind direction D sine value, wind direction D cosine is represented respectively Value, temperature T, air pressure P, specific volume a, the set of factors than wet H and roughness of ground surface R, here m=7;B, to n × 7 initial data (X) standardization of 7 set of factors, produced with eliminating the difference between variable on the order of magnitude Raw influence, the average for making each variable is 0, variance 1, obtains normalized matrix Y={ ynm};Wherein,And SmX is represented respectivelymAverage and variance.C, correlation matrix A is drawn according to normalized matrix;D, correlation matrix A characteristic value and characteristic vector are calculated, and calculates eigenvalue contribution rate and contribution rate of accumulative total;Wherein, λmRepresent A characteristic value.E, principal component Z=[Z are asked according to contribution rate of accumulative total1,Z2,...,Zm]Zm=YamWherein, amRepresent λmCharacteristic vector;m≤7.
- 4. a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis according to claim 1 or 3, its feature It is:The step (3) is by number of principal components according to the air speed data (v) in the n group initial data with the t-1 moment is collected The input matrix M of least square method supporting vector machine model is combined as, by the wind speed v of ttOutput matrix N as model Data are trained, finally obtain the tentative prediction sequence v'(t of wind speed);Wherein,In formula, dmnRepresent the data that principal component obtains.
- 5. a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis according to claim 1 or 3, its feature It is:The Lorenz equations are as follows:Wherein, x in formula, y, z are state variables, represent intensity, the horizontal direction temperature of convective motion fluid of convective motion respectively Departure degree when difference and vertical temperature-difference are to without convection current.σ, r, b are nondimensional arithmetic numbers, represent Prandtl number, auspicious respectively Sharp number and the amount related to climatic province size.For the parameter value of the Lorenz systems of classicsGiven initial value h =(1.1,1,1), make r=23, r=27, and we can obtain two groups of different numerical solutions, and then obtain two groups of different Lorenz Attractor.Because the numerical solution of Lorenz equations is one group of three-dimensional disturbance sequence, thus first non trivial solution is standardized Processing, obtain the standardization disturbance sequence of countless magnitudes and dimensionSpecific standardsization handle formula:X in formulan,yn,znRepresent the numerical solution of Lorenz equations, xmin,ymin,zminAnd xmax,ymax,zmaxX, y are represented respectively, z's Minimum value and maximum.
- 6. a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis according to claim 1 or 3, the step Suddenly the distance of second order Minkowski (Minkowski) described in (5) is as follows:Second order Minkowski Distance between normalized data and initial value (0,0,0):Wherein, CnRepresentC0Represent C0(x0,y0,z0)=C0(0,0,0)。
- 7. a kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis according to claim 1 or 3, its feature It is, using mean absolute error MAE and root-mean-square error RMSE come the quality of quantitative evaluation model, error amount is smaller, table The predictive ability of representation model is better.Its formula isWherein, y (t) and v (t) represents the actual value and predicted value of wind speed respectively, and k is forecast sample number.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710613264.8A CN107358006B (en) | 2017-07-25 | 2017-07-25 | Lorenz disturbance wind speed prediction method based on principal component analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710613264.8A CN107358006B (en) | 2017-07-25 | 2017-07-25 | Lorenz disturbance wind speed prediction method based on principal component analysis |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107358006A true CN107358006A (en) | 2017-11-17 |
CN107358006B CN107358006B (en) | 2021-10-22 |
Family
ID=60285412
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710613264.8A Active CN107358006B (en) | 2017-07-25 | 2017-07-25 | Lorenz disturbance wind speed prediction method based on principal component analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107358006B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108846508A (en) * | 2018-05-30 | 2018-11-20 | 华北电力大学(保定) | A kind of wind speed forecasting method and system based on atmospheric perturbation |
CN109991683A (en) * | 2019-04-12 | 2019-07-09 | 中国气象局沈阳大气环境研究所 | Wind frequency is according to processing method and processing device |
CN110687255A (en) * | 2019-10-21 | 2020-01-14 | 软通动力信息技术有限公司 | Air pollutant tracing method, device, equipment and storage medium |
CN112464154A (en) * | 2020-11-27 | 2021-03-09 | 中国船舶重工集团公司第七0四研究所 | Method for automatically screening effective features based on unsupervised learning |
CN112926758A (en) * | 2019-12-05 | 2021-06-08 | 中国科学院深圳先进技术研究院 | Short-term wind speed prediction method and system |
CN114675055A (en) * | 2022-03-29 | 2022-06-28 | 北京航空航天大学 | Air speed measurement error compensation method for sonde based on inertial system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104504466A (en) * | 2014-12-17 | 2015-04-08 | 华北电力大学(保定) | Wind power plant power prediction method considering atmospheric disturbance effect |
CN104657584A (en) * | 2014-12-16 | 2015-05-27 | 华北电力大学(保定) | Lorenz-system-based wind speed prediction method |
CN104933469A (en) * | 2015-03-02 | 2015-09-23 | 华北电力大学(保定) | Short-term wind speed forecasting method based on grey generating perturbation model |
CN106384170A (en) * | 2016-09-24 | 2017-02-08 | 华北电力大学(保定) | Wavelet decomposition and reconstruction-based time sequence wind speed prediction method |
-
2017
- 2017-07-25 CN CN201710613264.8A patent/CN107358006B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104657584A (en) * | 2014-12-16 | 2015-05-27 | 华北电力大学(保定) | Lorenz-system-based wind speed prediction method |
CN104504466A (en) * | 2014-12-17 | 2015-04-08 | 华北电力大学(保定) | Wind power plant power prediction method considering atmospheric disturbance effect |
CN104933469A (en) * | 2015-03-02 | 2015-09-23 | 华北电力大学(保定) | Short-term wind speed forecasting method based on grey generating perturbation model |
CN106384170A (en) * | 2016-09-24 | 2017-02-08 | 华北电力大学(保定) | Wavelet decomposition and reconstruction-based time sequence wind speed prediction method |
Non-Patent Citations (1)
Title |
---|
YAGANG ZHANG等: "New Progress in Wind Prediction Based on Nonlinear Amendment", 《2014 IEEE FOURTH INTERNATIONAL CONFERENCE ON BIG DATA AND CLOUD COMPUTING》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108846508A (en) * | 2018-05-30 | 2018-11-20 | 华北电力大学(保定) | A kind of wind speed forecasting method and system based on atmospheric perturbation |
CN109991683A (en) * | 2019-04-12 | 2019-07-09 | 中国气象局沈阳大气环境研究所 | Wind frequency is according to processing method and processing device |
CN110687255A (en) * | 2019-10-21 | 2020-01-14 | 软通动力信息技术有限公司 | Air pollutant tracing method, device, equipment and storage medium |
CN112926758A (en) * | 2019-12-05 | 2021-06-08 | 中国科学院深圳先进技术研究院 | Short-term wind speed prediction method and system |
CN112464154A (en) * | 2020-11-27 | 2021-03-09 | 中国船舶重工集团公司第七0四研究所 | Method for automatically screening effective features based on unsupervised learning |
CN112464154B (en) * | 2020-11-27 | 2024-03-01 | 中国船舶重工集团公司第七0四研究所 | Method for automatically screening effective features based on unsupervised learning |
CN114675055A (en) * | 2022-03-29 | 2022-06-28 | 北京航空航天大学 | Air speed measurement error compensation method for sonde based on inertial system |
Also Published As
Publication number | Publication date |
---|---|
CN107358006B (en) | 2021-10-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107358006A (en) | A kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis | |
CN103268366B (en) | A kind of combination wind power forecasting method suitable for distributing wind power plant | |
CN105046374B (en) | A kind of power interval prediction technique based on core extreme learning machine model | |
CN104732300B (en) | A kind of neutral net wind power short term prediction method theoretical based on Fuzzy divide | |
CN109086928A (en) | Photovoltaic plant realtime power prediction technique based on SAGA-FCM-LSSVM model | |
CN109086534A (en) | A kind of wind power plant wake flow correction method and system based on CFD fluid mechanic model | |
CN105389634A (en) | Combined short-term wind power prediction system and method | |
CN104933483A (en) | Wind power forecasting method dividing based on weather process | |
CN104036135B (en) | A kind of variational Assimilation method under typhoon dynamic equilibrium constraint based on WRF patterns | |
CN107276069A (en) | Approximate the polymerization modeling method and system of a kind of area power grid resident temperature control load | |
CN103617308B (en) | A kind of construction method of wind power plant frequency domain equivalent model | |
CN115345076B (en) | Wind speed correction processing method and device | |
CN105447572A (en) | Wind power prediction system and method based on neural network optimized by genetic algorithm | |
CN109146192B (en) | Wind power prediction method considering operation condition of wind turbine generator | |
CN110991701A (en) | Wind power plant fan wind speed prediction method and system based on data fusion | |
CN104657584A (en) | Lorenz-system-based wind speed prediction method | |
Zhang et al. | Improved wind prediction based on the Lorenz system | |
CN105023099B (en) | A kind of wind-driven generator output appraisal procedure for considering turbulence intensity | |
CN103699804B (en) | Method for monitoring power output deviation in real time during running of wind turbines | |
CN108536907B (en) | Wind turbine far-field wake flow analytic modeling method based on simplified momentum theorem | |
CN104700156B (en) | A kind of wind power forecasting method based on support vector machines selection variables | |
CN106779202B (en) | Wind power prediction method considering air humidity | |
CN115034159A (en) | Power prediction method, device, storage medium and system for offshore wind farm | |
CN114943174A (en) | Fan output loss prediction method used under cold tide small sample condition | |
CN103696913A (en) | Method for monitoring wind power utilization deviation in real time during running of wind turbines |
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 |