CN108846508A - A kind of wind speed forecasting method and system based on atmospheric perturbation - Google Patents
A kind of wind speed forecasting method and system based on atmospheric perturbation Download PDFInfo
- Publication number
- CN108846508A CN108846508A CN201810537126.0A CN201810537126A CN108846508A CN 108846508 A CN108846508 A CN 108846508A CN 201810537126 A CN201810537126 A CN 201810537126A CN 108846508 A CN108846508 A CN 108846508A
- Authority
- CN
- China
- Prior art keywords
- wind speed
- sequence
- forecasting
- value
- disturbed
- 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
- 238000013277 forecasting method Methods 0.000 title claims abstract description 17
- 230000015654 memory Effects 0.000 claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims description 16
- 241001269238 Data Species 0.000 claims description 6
- 239000012530 fluid Substances 0.000 claims description 6
- 230000006403 short-term memory Effects 0.000 claims description 2
- 230000008439 repair process Effects 0.000 claims 1
- 238000013528 artificial neural network Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000005611 electricity Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012163 sequencing technique Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000005183 dynamical system Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 101150040378 steep1 gene Proteins 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000003313 weakening 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Biophysics (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of wind speed forecasting method and system based on atmospheric perturbation.The prediction technique includes：Obtain the trained forecasting wind speed model based on long memory network in short-term；The tentative prediction sequence of wind speed is obtained according to the forecasting wind speed model, the corresponding wind speed initial predicted value of each moment in the tentative prediction sequence；Obtain the primary condition and parameter of lorentz equation；Sequence is disturbed according to the primary condition and the gain of parameter wind speed, the wind speed disturbance sequence is the numerical solution of the lorentz equation, and the wind speed disturbance sequence includes corresponding wind speed disturbed value of each moment；Sequence is disturbed according to the wind speed to be modified the tentative prediction sequence, obtains the final forecasting sequence of wind speed.Using wind speed forecasting method or system of the invention, prediction result has smaller prediction error closer to true wind series.
Description
Technical field
The present invention relates to forecasting wind speed fields, more particularly to a kind of wind speed forecasting method based on atmospheric perturbation and are
System.
Background technique
Since twentieth century end, being continuously increased for world energy consumption causes global climate persistently to warm, while into one
Step destroys the balance of natural ecosystems, threatens the food supply and living environment of the mankind.According to international energy on March 22nd, 2018
The data of source administration publication show that global carbon emission amount was up to 32,500,000,000 tons in 2017, increase by 1.4% than 2016 and (have a net increase of long 4.6
Hundred million tons), the main reason for recording high, leading to this status is that global energy requirements are powerful.Currently, development and utilization are renewable
The energy has become the strategic choice of most countries in the world, and many countries are using Renewable Energy Development as alleviation energy supply
Contradiction, the important measures for coping with climate change.Wherein wind-power electricity generation as it is a kind of without fuel consumption and operating cost it is lower
Generation mode becomes the object that every country falls over each other research.Since very significant change may occur within a few hours for wind energy,
So the problem of wind-power electricity generation maximum is that its dependence to wind energy fluctuation, it is this to rely on the fluctuation for further resulting in wind-powered electricity generation
Property, it is unfavorable for the grid-connected utilization of wind-powered electricity generation.Therefore, accurate forecasting wind speed be the key that large-scale develop and utilize wind energy resources, but
It is simultaneously also very uppity link.
In recent years, researcher both domestic and external propose a variety of different compared with in-depth study to forecasting wind speed
Wind speed forecasting method, such as the wind speed forecasting method based on meteorological data, the wind speed forecasting method based on least square method are based on
The prediction technique of statistical regression analysis and neural network, the wind speed forecasting method etc. based on intelligent algorithm, but they are almost
It all rests in the improvement to algorithm, and has ignored nonlinear characteristic of the wind energy in big pneumatic power system.Therefore, traditional
In wind speed forecasting method, the precision generally predicted is not high.
Summary of the invention
The object of the present invention is to provide a kind of wind speed forecasting method and system based on atmospheric perturbation, to improve forecasting wind speed
Precision.
To achieve the above object, the present invention provides following schemes：
A kind of wind speed forecasting method based on atmospheric perturbation, the prediction technique include：
Obtain the trained forecasting wind speed model based on long memory network in short-term；
The tentative prediction sequence of wind speed, each moment in the tentative prediction sequence are obtained according to the forecasting wind speed model
A corresponding wind speed initial predicted value；
Obtain the primary condition and parameter of lorentz equation；
Sequence is disturbed according to the primary condition and the gain of parameter wind speed, the wind speed disturbance sequence is the long-range navigation
The hereby numerical solution of equation, the wind speed disturbance sequence includes corresponding wind speed disturbed value of each moment；
Sequence is disturbed according to the wind speed to be modified the tentative prediction sequence, obtains the final pre- sequencing of wind speed
Column.
Optionally, described to obtain the trained forecasting wind speed model based on long memory network in short-term, further include before：
Obtain the forecasting wind speed model based on long memory network in short-term；
Using 800 air speed datas as the input of network training collection, using 100 wind speed values as network training collection
Output, is trained the forecasting wind speed model.
Optionally, the primary condition of the lorentz equation is：The initial value h=(1.1,1,1) of (x, y, z)；X indicates convection current
The amplitude of movement, y indicate the horizontal direction temperature difference of rise and fall fluid when convection current, and z indicates vertical direction temperature caused by convection current
Deviation of the difference to linear case；
The parameter of the lorentz equation is：σ=10, b=8/3, r=8, wherein σ is Prandtl number and r is Rayleigh
Number, b are parameter related with container size shape.
Optionally, the numerical solution of the lorentz equation according to the primary condition and the gain of parameter, it is described
Numerical solution is that the wind speed of each moment wind speed disturbed value composition disturbs sequence, further includes later：
It utilizesTo each moment
Wind speed disturbed value be standardized, obtain standardized wind speed disturbed value, and then obtain standardized wind speed disturbance sequence
Column；Wherein (xn,yn,zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xmin
For the minimum value of x in all wind speed disturbed values, yminFor the minimum value of y in all wind speed disturbed values, zminFor the disturbance of all wind speed
The minimum value of z, x in valuemaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values,
zmaxFor the maximum value of z in all wind speed disturbed values.
Optionally, described that the tentative prediction sequence is modified according to wind speed disturbance sequence, obtain wind speed
Final forecasting sequence, specifically includes：
It utilizesIt determines n-th in the wind speed disturbance sequence
Moment corresponding strength of turbulence d (tn), and then obtain wind speed strength of turbulence sequence d (t1,t2,...,tk)；Wherein
For the wind speed disturbed value at the n-th moment, (x0,y0,z0) be wind speed disturbed value initial value；
Utilize v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to the tentative prediction sequence
Column are modified, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,tk) it is revised
The final forecasting sequence of wind speed, wherein the final predicted value of the n-th moment revised wind speed is v " (tn), l is coefficient of disturbance.
The forecasting wind speed system based on atmospheric perturbation that the present invention also provides a kind of, the forecasting system include：
Forecasting wind speed model obtains module, for obtaining the trained forecasting wind speed mould based on long memory network in short-term
Type；
Tentative prediction retrieval module, for obtaining the tentative prediction sequence of wind speed according to the forecasting wind speed model,
The corresponding wind speed initial predicted value of each moment in the tentative prediction sequence；
Primary condition and parameter acquisition module, for obtaining the primary condition and parameter of lorentz equation；
Wind speed disturbs retrieval module, for disturbing sequence according to the primary condition and the gain of parameter wind speed,
The wind speed disturbance sequence is the numerical solution of the lorentz equation, and the wind speed disturbance sequence includes corresponding wind of each moment
Fast disturbed value；
The final forecasting sequence of wind speed obtains module, for according to the wind speed disturb sequence to the tentative prediction sequence into
Row amendment, obtains the final forecasting sequence of wind speed.
Optionally, the system also includes：
Initial forecasting wind speed model obtains module, for obtaining the trained forecasting wind speed based on long memory network in short-term
Before model, the forecasting wind speed model based on long memory network in short-term is obtained；
Training module, for using 800 air speed datas as the input of network training collection, 100 wind speed values to be made
For the output of network training collection, the forecasting wind speed model is trained.
Optionally, the primary condition of the lorentz equation is：The initial value h=(1.1,1,1) of (x, y, z)；X indicates convection current
The amplitude of movement, y indicate the horizontal direction temperature difference of rise and fall fluid when convection current, and z indicates vertical direction temperature caused by convection current
Deviation of the difference to linear case；
The parameter of the lorentz equation is：σ=10, b=8/3, r=8, wherein σ is Prandtl number and r is Rayleigh
Number, b are parameter related with container size shape.
Optionally, the system also includes：
Standardized module utilizes after according to the primary condition and gain of parameter wind speed disturbance sequenceTo the wind speed disturbed value at each moment into
Row standardization obtains standardized wind speed disturbed value, and then obtains standardized wind speed disturbance sequence；Wherein (xn,yn,
zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xminFor the disturbance of all wind speed
The minimum value of x, y in valueminFor the minimum value of y in all wind speed disturbed values, zminFor the minimum value of z in all wind speed disturbed values,
xmaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values, zmaxIt is disturbed for all wind speed
The maximum value of z in dynamic value.
Optionally, the final forecasting sequence of the wind speed obtains module, specifically includes：
Strength of turbulence acquiring unit, for utilizingIt determines
N-th moment corresponding strength of turbulence d (t in the wind speed disturbance sequencen), and then obtain wind speed strength of turbulence sequence d (t1,
t2,...,tk)；WhereinFor the wind speed disturbed value at the n-th moment, (x0,y0,z0) be wind speed disturbed value initial value；
Amending unit, for utilizing v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to institute
It states tentative prediction sequence to be modified, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,
tk) be revised wind speed final forecasting sequence, wherein the final predicted value of the n-th moment revised wind speed be v " (tn), l
For coefficient of disturbance.
The specific embodiment provided according to the present invention, the invention discloses following technical effects：
Wind speed forecasting method and system of the invention is that the Lorentz based on long memory network in short-term disturbs forecasting wind speed mistake
Journey carries out preliminary prediction to original wind series with long memory network in short-term first, secondly considers the chaos fortune of Atmosphere System
It is dynamic, the initial predicted value of wind series is modified with Lorentz disturbance sequence, can more accurately describe wind speed in this way
Actual change rule, weaken the randomness and fluctuation of wind series.The present invention is based on the long-range navigations of long memory network in short-term
Hereby disturbance wind speed forecasting method has better prediction effect, and prediction result has more closer to true wind series
Small prediction error.Therefore, during big pneumatic power system is integrated to forecasting wind speed by prediction technique of the invention and system,
The precision for improving forecasting wind speed may advantageously facilitate the high speed development of Wind Power Generation Industry.
Detailed description of the invention
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the flow diagram of wind speed forecasting method of the present invention；
Fig. 2 is the structural schematic diagram of forecasting wind speed system of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing and specific real
Applying mode, the present invention is described in further detail.
Air motion is a deterministic dynamical system, and dynamic behavior can use one group of Lorentz (Lorenz) side
Journey describes.Meanwhile long memory network (LSTM in short-term：Long Short-Term Memory) it is a kind of time recurrent neural net
Network, its difference from traditional feedforward neural network (feed-forward network) are that LSTM can be defeated to before
Enter content selectively to remember, to help to judge current input.This feature of LSTM is related in processing time series
Input when have very big advantage, suitable for processing and predicted time sequence in be spaced and postpone relatively long critical event,
And wind series are exactly a kind of and former longer period of time relies on close time series.Therefore, the present invention is used based on LSTM mind
Lorenz disturbance forecasting wind speed model through network predicts wind speed.Firstly, being carried out tentatively to raw data set with LSTM
Prediction, obtains preliminary forecasting wind speed model；Meanwhile it solving Lorenz equation and obtaining atmospheric perturbation sequence, simulated atmosphere movement
Process, with the tentative prediction of Lorenz disturbance sequence amendment wind speed as a result, improving tentative prediction as a result, weakening wind speed with this
Randomness and uncertainty make the prediction process of wind speed more meet its actual movement rule, and then improve the precision of prediction of wind speed.
Fig. 1 is the flow diagram of wind speed forecasting method of the present invention.As shown in Figure 1, the prediction technique includes：
Step 100：Obtain the trained forecasting wind speed model based on long memory network in short-term.LSTM neural network can be very
The characteristics of grasp wind speed time series got well, excavates the regular air speed value to predict subsequent time from past air speed value, because
This, carries out forecasting wind speed using the forecasting wind speed model based on LSTM neural network.
LSTM is a kind of neural network, and LSTM neural network is a kind of special shape of RNN neural network.General RNN
Dependence Problem when greatest problem existing for neural network is long, i.e. RNN can not learn to it is being relied on future position but at a distance of compared with
The relevant information of remote point.But at present it has been proved that LSTM neural network be solve RNN neural network it is long when Dependence Problem have
Effect technology, and this technology has very high universality, a possibility that bringing variation is very more, and this is further such that LSTM
It can handle ever-changing Perpendicular Problems.LSTM is different from the place of common RNN, and essentially consist in it joined one in the algorithm
The structure of a " processor " judged whether information is useful, the effect of this processor is referred to as cell.It is put in one cell
Three fan doors have been set, has been called input gate respectively, forgets door and out gate, range is between 0 to 1.One information enters the net of LSTM
, can be according to rule to determine whether useful in network, the information for only meeting algorithm certification can just leave, and the information not being inconsistent is then
It is passed into silence by forgeing door.It is in simple terms exactly the working principle of one-in-and-two-out, such structure can effectively mitigate disappearance
Gradient problem and undated parameter can be learnt automatically.The treatment process of each cell is as follows：
Steep1：It is removed, is forgotten using which type of information of " forgeing thresholding layer " to determine in cell state
The activation primitive f of doort, ft=σ (Wf·[ht-1,xt]+bf)；
Steep2：Determine which type of new information should be stored in cell state, firstly, utilizing " input threshold layer "
Sigmoid layer determine which value need to update, obtain input threshold function it, then utilize one packet of a tanh layers of creation
Containing the vector that can be added to candidate value new in this stateWherein, it=σ (Wf·[ht-1,xt]+bi),
Steep3：Cell state is updated from Ct-1To Ct,
Steep4：Selection output, first in sigmoid layers of determining means state which partially need to export, obtain
Export threshold function ot, then location mode is input in tanh function (value is converted between -1 to 1) multiplied by output
Sigmoid threshold value obtain the output h of active cellt, ot=σ (Wo·[ht-1,xt]+bo), ht=ot*tanh(Ct), wherein
Wi,Wf,Wc,WoRepresent weight matrix, bi,bf,bc,boRepresent corresponding bias vector, ht-1The output of previous moment unit is represented,
xtRepresent input next time.After constructing the forecasting wind speed model based on LSTM, need to carry out forecasting wind speed model
Training, using 800 air speed datas as the input of network training collection, using 100 wind speed values as the defeated of network training collection
Out, the forecasting wind speed model is trained.
Step 200：The tentative prediction sequence of wind speed is obtained according to forecasting wind speed model.It is pre- using the wind speed based on LSTM
It surveys model and carries out forecasting wind speed, available each moment corresponding wind speed initial predicted value is available to press after the completion of prediction
The tentative prediction sequence of all wind speed initial predicted values composition of time-sequencing.
Step 300：Obtain the primary condition and parameter of lorentz equation.Since small disturbance any in atmosphere all may
Tremendous influence is caused to Atmosphere System.So considering that small sample perturbations are intentional when studying the wind speed variation in Atmosphere System
It is adopted and necessary.The present invention measures the disturbance of atmosphere using the Disturbance Model of lorentz equation building, therefore, first
It needs to be determined that the primary condition and parameter of lorentz equation.The parameter of lorentz equation is in the present invention：σ=10, b=8/3, r
=8, wherein σ is Prandtl number and r is Rayleigh number, and b is parameter related with container size shape.
It is best according to prediction effect is disturbed when testing and determining that (x, y, z) initial value is (1.1,1,1), therefore, by Lorentz side
The primary condition of journey is determined as：The initial value h=(1.1,1,1) of (x, y, z)；X indicates the amplitude of convective motion, when y indicates convection current
The horizontal direction temperature difference of rise and fall fluid, z indicate deviation of the vertical direction temperature difference to linear case caused by convection current.
Step 400：Sequence is disturbed according to primary condition and gain of parameter wind speed.The wind speed disturbance sequence is the long-range navigation
The hereby numerical solution of equation, the wind speed disturbance sequence includes corresponding wind speed disturbed value of each moment.
Bring initial value and parameter into lorentz equation firstIn, and then determine x (t), y (t), z
(t) function expression；Then the specific value of x (t), y (t), z (t) are determined according to the specific moment, it is corresponding in conjunction with the moment
Wind speed initial predicted value determines that the numerical solution of the moment corresponding lorentz equation, the numerical solution at the moment are wind speed disturbance
Value.Wind speed is obtained according to the wind speed disturbed value at each moment and disturbs sequence, and then obtains its Lorenz attractor.
Since the numerical solution of Lorenz equation is one group of three-dimensional disturbance sequence, thus needs to disturb sequence to wind speed and carry out
Standardization, obtains the standardization disturbance sequence of countless magnitudes and dimension, and specific course of standardization process is as follows：
It utilizesTo each moment
Wind speed disturbed value is standardized, and obtains standardized wind speed disturbed value, and then obtains standardized wind speed disturbance sequence；
Wherein (xn,yn,zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xminFor
The minimum value of x, y in all wind speed disturbed valuesminFor the minimum value of y in all wind speed disturbed values, zminFor all wind speed disturbed values
The minimum value of middle z, xmaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values, zmax
For the maximum value of z in all wind speed disturbed values.
Step 500：Sequence is disturbed according to wind speed to be modified tentative prediction sequence, obtains the final pre- sequencing of wind speed
Column.Solution vector when chaotic motion state occurs in present invention application Lorenz system is disturbed to construct disturbance sequence by what is obtained
Dynamic sequence acts in tentative prediction result, obtains the improvement to tentative prediction result.Detailed process is as follows：
It utilizesIt determines n-th in the wind speed disturbance sequence
Moment corresponding strength of turbulence d (tn), and then obtain wind speed strength of turbulence sequence d (t1,t2,...,tk)；Wherein
For the wind speed disturbed value at the n-th moment, (x0,y0,z0) be wind speed disturbed value initial value；
Utilize v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to the tentative prediction sequence
Column are modified, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,tk) it is revised
The final forecasting sequence of wind speed, wherein the final predicted value of the n-th moment revised wind speed is v " (tn), l is coefficient of disturbance.If
L is positive number, then shows that wind speed tentative prediction sequence needs to reinforce disturbance；If l is negative, show that wind speed tentative prediction sequence needs
Weaken disturbance；K is the sample number of prediction, the i.e. sample number that the k moment obtains.
Fig. 2 is the structural schematic diagram of forecasting wind speed system of the present invention.As shown in Fig. 2, the forecasting system includes：
Forecasting wind speed model obtains module 201, for obtaining the trained forecasting wind speed based on long memory network in short-term
Model.
Tentative prediction retrieval module 202, for obtaining the tentative prediction sequence of wind speed according to the forecasting wind speed model
It arranges, the corresponding wind speed initial predicted value of each moment in the tentative prediction sequence.
Primary condition and parameter acquisition module 203, for obtaining the primary condition and parameter of lorentz equation.The long-range navigation
Hereby the primary condition of equation is：The initial value h=(1.1,1,1) of (x, y, z)；X indicates the amplitude of convective motion, when y indicates convection current
The horizontal direction temperature difference of rise and fall fluid, z indicate deviation of the vertical direction temperature difference to linear case caused by convection current；It is described
The parameter of lorentz equation is：σ=10, b=8/3, r=8, wherein σ is Prandtl number and r is Rayleigh number, and b is big with container
The related parameter of small shape.
Wind speed disturbs retrieval module 204, for disturbing sequence according to the primary condition and the gain of parameter wind speed
Column, the wind speed disturbance sequence are the numerical solution of the lorentz equation, and the wind speed disturbance sequence includes corresponding at each moment
Wind speed disturbed value.
The final forecasting sequence of wind speed obtains module 205, for disturbing sequence to the tentative prediction sequence according to the wind speed
Column are modified, and obtain the final forecasting sequence of wind speed.
Wherein, the system also includes：
Initial forecasting wind speed model obtains module, for obtaining the trained forecasting wind speed based on long memory network in short-term
Before model, the forecasting wind speed model based on long memory network in short-term is obtained；
Training module, for using 800 air speed datas as the input of network training collection, 100 wind speed values to be made
For the output of network training collection, the forecasting wind speed model is trained.
The system also includes：
Standardized module utilizes after according to the primary condition and gain of parameter wind speed disturbance sequenceTo the wind speed disturbed value at each moment
It is standardized, obtains standardized wind speed disturbed value, and then obtain standardized wind speed disturbance sequence；Wherein (xn,yn,
zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xminFor the disturbance of all wind speed
The minimum value of x, y in valueminFor the minimum value of y in all wind speed disturbed values, zminFor the minimum value of z in all wind speed disturbed values,
xmaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values, zmaxIt is disturbed for all wind speed
The maximum value of z in dynamic value.
The final forecasting sequence of wind speed obtains module 205, specifically includes：
Strength of turbulence acquiring unit, for utilizingIt determines
N-th moment corresponding strength of turbulence d (t in the wind speed disturbance sequencen), and then obtain wind speed strength of turbulence sequence d (t1,
t2,...,tk)；WhereinFor the wind speed disturbed value at the n-th moment, (x0,y0,z0) be wind speed disturbed value initial value；
Amending unit, for utilizing v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to institute
It states tentative prediction sequence to be modified, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,
tk) be revised wind speed final forecasting sequence, wherein the final predicted value of the n-th moment revised wind speed be v " (tn), l
For coefficient of disturbance.
The present invention carries out disturbance amendment to the tentative prediction sequence of wind speed using the comprehensive disturbance stream of Lorenz, and to realize wind
The minimum target of mean absolute error between fast actual value and predicted value obtains disturbing modified Optimal Disturbance intensity and optimal
Coefficient of disturbance；Tentative prediction of the Optimal Disturbance intensity and Optimal Disturbance coefficient obtained using the comprehensive disturbance stream of Lorenz to wind speed
Sequence carries out disturbance amendment, obtains the disturbance Orders Corrected of wind speed, the precision of prediction is higher, and error is smaller.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For system disclosed in embodiment
For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part
It is bright.
Used herein a specific example illustrates the principle and implementation of the invention, and above embodiments are said
It is bright to be merely used to help understand method and its core concept of the invention；At the same time, for those skilled in the art, foundation
Thought of the invention, there will be changes in the specific implementation manner and application range.In conclusion the content of the present specification is not
It is interpreted as limitation of the present invention.
Claims (10)
1. a kind of wind speed forecasting method based on atmospheric perturbation, which is characterized in that the prediction technique includes：
Obtain the trained forecasting wind speed model based on long memory network in short-term；
The tentative prediction sequence of wind speed is obtained according to the forecasting wind speed model, each moment is corresponding in the tentative prediction sequence
One wind speed initial predicted value；
Obtain the primary condition and parameter of lorentz equation；
Sequence is disturbed according to the primary condition and the gain of parameter wind speed, the wind speed disturbance sequence is the Lorentz side
The numerical solution of journey, the wind speed disturbance sequence includes corresponding wind speed disturbed value of each moment；
Sequence is disturbed according to the wind speed to be modified the tentative prediction sequence, obtains the final forecasting sequence of wind speed.
2. prediction technique according to claim 1, which is characterized in that the acquisition is trained to be based on long short-term memory net
The forecasting wind speed model of network further includes before：
Obtain the forecasting wind speed model based on long memory network in short-term；
Using 800 air speed datas as the input of network training collection, using 100 wind speed values as the defeated of network training collection
Out, the forecasting wind speed model is trained.
3. prediction technique according to claim 1, which is characterized in that the primary condition of the lorentz equation is：(x,y,
Z) initial value h=(1.1,1,1)；X indicates that the amplitude of convective motion, y indicate the horizontal direction of rise and fall fluid when convection current
The temperature difference, z indicate deviation of the vertical direction temperature difference to linear case caused by convection current；
The parameter of the lorentz equation is：σ=10, b=8/3, r=8, wherein σ is Prandtl number and r is Rayleigh number, and b is
Parameter related with container size shape.
4. prediction technique according to claim 3, which is characterized in that described to be obtained according to the primary condition and the parameter
The numerical solution of the lorentz equation is obtained, the numerical solution is that the wind speed of each moment wind speed disturbed value composition disturbs sequence, it
After further include：
It utilizesTo the wind speed at each moment
Disturbed value is standardized, and obtains standardized wind speed disturbed value, and then obtains standardized wind speed disturbance sequence；Wherein
(xn,yn,zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xminIt is all
The minimum value of x, y in wind speed disturbed valueminFor the minimum value of y in all wind speed disturbed values, zminFor z in all wind speed disturbed values
Minimum value, xmaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values, zmaxFor institute
There is the maximum value of z in wind speed disturbed value.
5. prediction technique according to claim 4, which is characterized in that described to disturb sequence to described first according to the wind speed
Step forecasting sequence is modified, and is obtained the final forecasting sequence of wind speed, is specifically included：
It utilizesDetermine the n-th moment in the wind speed disturbance sequence
Corresponding strength of turbulence d (tn), and then obtain wind speed strength of turbulence sequence d (t1,t2,...,tk)；WhereinIt is
The wind speed disturbed value at n moment, (x0,y0,z0) be wind speed disturbed value initial value；
Utilize v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to the tentative prediction sequence into
Row is corrected, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,tk) it is revised wind speed
Final forecasting sequence, wherein the final predicted value of the n-th moment revised wind speed be v " (tn), l is coefficient of disturbance.
6. a kind of forecasting wind speed system based on atmospheric perturbation, which is characterized in that the forecasting system includes：
Forecasting wind speed model obtains module, for obtaining the trained forecasting wind speed model based on long memory network in short-term；
Tentative prediction retrieval module, it is described for obtaining the tentative prediction sequence of wind speed according to the forecasting wind speed model
The corresponding wind speed initial predicted value of each moment in tentative prediction sequence；
Primary condition and parameter acquisition module, for obtaining the primary condition and parameter of lorentz equation；
Wind speed disturbs retrieval module, described for disturbing sequence according to the primary condition and the gain of parameter wind speed
Wind speed disturbs the numerical solution that sequence is the lorentz equation, and the wind speed disturbance sequence includes that corresponding wind speed of each moment is disturbed
Dynamic value；
The final forecasting sequence of wind speed obtains module, repairs for disturbing sequence according to the wind speed to the tentative prediction sequence
Just, the final forecasting sequence of wind speed is obtained.
7. forecasting system according to claim 6, which is characterized in that the system also includes：
Initial forecasting wind speed model obtains module, for obtaining the trained forecasting wind speed model based on long memory network in short-term
Before, the forecasting wind speed model based on long memory network in short-term is obtained；
Training module, for using 800 air speed datas as the input of network training collection, using 100 wind speed values as net
The output of network training set is trained the forecasting wind speed model.
8. forecasting system according to claim 6, which is characterized in that the primary condition of the lorentz equation is：(x,y,
Z) initial value h=(1.1,1,1)；X indicates that the amplitude of convective motion, y indicate the horizontal direction of rise and fall fluid when convection current
The temperature difference, z indicate deviation of the vertical direction temperature difference to linear case caused by convection current；
The parameter of the lorentz equation is：σ=10, b=8/3, r=8, wherein σ is Prandtl number and r is Rayleigh number, and b is
Parameter related with container size shape.
9. forecasting system according to claim 8, which is characterized in that the system also includes：
Standardized module utilizes after according to the primary condition and gain of parameter wind speed disturbance sequenceTo the wind speed disturbed value at each moment
It is standardized, obtains standardized wind speed disturbed value, and then obtain standardized wind speed disturbance sequence；Wherein (xn,yn,
zn) be the n moment wind speed disturbed value,For the standardized wind speed disturbed value at n moment；xminFor the disturbance of all wind speed
The minimum value of x, y in valueminFor the minimum value of y in all wind speed disturbed values, zminFor the minimum value of z in all wind speed disturbed values,
xmaxFor the maximum value of x in all wind speed disturbed values, ymaxFor the maximum value of y in all wind speed disturbed values, zmaxIt is disturbed for all wind speed
The maximum value of z in dynamic value.
10. forecasting system according to claim 9, which is characterized in that the final forecasting sequence of wind speed obtains module, tool
Body includes：
Strength of turbulence acquiring unit, for utilizingDescribed in determination
Wind speed disturbs the n-th moment corresponding strength of turbulence d (t in sequencen), and then obtain wind speed strength of turbulence sequence d (t1,t2,...,
tk)；WhereinFor the wind speed disturbed value at the n-th moment, (x0,y0,z0) be wind speed disturbed value initial value；
Amending unit, for utilizing v " (t1,t2,...,tk)=v'(t1,t2,...,tk)+ld(t1,t2,...,tk) to described first
Step forecasting sequence is modified, wherein v'(t1,t2,...,tk) be wind speed tentative prediction sequence, v " (t1,t2,...,tk) be
The final forecasting sequence of revised wind speed, wherein the final predicted value of the n-th moment revised wind speed is v " (tn), l is to disturb
Dynamic coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810537126.0A CN108846508A (en) | 2018-05-30 | 2018-05-30 | A kind of wind speed forecasting method and system based on atmospheric perturbation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810537126.0A CN108846508A (en) | 2018-05-30 | 2018-05-30 | A kind of wind speed forecasting method and system based on atmospheric perturbation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108846508A true CN108846508A (en) | 2018-11-20 |
Family
ID=64210012
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810537126.0A Pending CN108846508A (en) | 2018-05-30 | 2018-05-30 | A kind of wind speed forecasting method and system based on atmospheric perturbation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108846508A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060173623A1 (en) * | 2005-02-01 | 2006-08-03 | Grzych Matthew L | System and method for enhanced measure-correlate-predict for a wind farm location |
| CN103400052A (en) * | 2013-08-22 | 2013-11-20 | 武汉大学 | Combined method for predicting short-term wind speed in wind power plant |
| US20140172329A1 (en) * | 2012-12-17 | 2014-06-19 | Junshan Zhang | System and method for wind generation forecasting |
| 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 |
| CN107358006A (en) * | 2017-07-25 | 2017-11-17 | 华北电力大学（保定） | A kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis |
| CN107392379A (en) * | 2017-07-25 | 2017-11-24 | 华北电力大学（保定） | A kind of time series wind speed forecasting method based on Lorenz disturbances |
-
2018
- 2018-05-30 CN CN201810537126.0A patent/CN108846508A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060173623A1 (en) * | 2005-02-01 | 2006-08-03 | Grzych Matthew L | System and method for enhanced measure-correlate-predict for a wind farm location |
| US20140172329A1 (en) * | 2012-12-17 | 2014-06-19 | Junshan Zhang | System and method for wind generation forecasting |
| CN103400052A (en) * | 2013-08-22 | 2013-11-20 | 武汉大学 | Combined method for predicting short-term wind speed in wind power plant |
| 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 |
| CN107358006A (en) * | 2017-07-25 | 2017-11-17 | 华北电力大学（保定） | A kind of Lorenz disturbance wind speed forecasting methods based on principal component analysis |
| CN107392379A (en) * | 2017-07-25 | 2017-11-24 | 华北电力大学（保定） | A kind of time series wind speed forecasting method based on Lorenz disturbances |
Non-Patent Citations (1)
| Title |
|---|
| 姚万业 等: ""基于风速预测的改进爬山法最大功率追踪策略"", 《电力科学与工程》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111179121B (en) | Power grid emergency control method based on expert system and deep reverse reinforcement learning | |
| CN104376389B (en) | Master-slave mode microgrid power load prediction system and method based on load balancing | |
| CN103226741B (en) | Public supply mains tube explosion prediction method | |
| CN105354363A (en) | Fluctuation wind speed prediction method based on extreme learning machine | |
| CN110942205A (en) | Short-term photovoltaic power generation power prediction method based on HIMVO-SVM | |
| CN111695724B (en) | Wind speed prediction method based on hybrid neural network model | |
| CN107121926A (en) | A kind of industrial robot Reliability Modeling based on deep learning | |
| CN111415010A (en) | Bayesian neural network-based wind turbine generator parameter identification method | |
| CN105243259A (en) | Extreme learning machine based rapid prediction method for fluctuating wind speed | |
| CN103927460A (en) | Wind power plant short-term wind speed prediction method based on RBF | |
| CN109255726A (en) | A kind of ultra-short term wind power prediction method of Hybrid Intelligent Technology | |
| CN114119273A (en) | Park comprehensive energy system non-invasive load decomposition method and system | |
| CN108832663A (en) | The prediction technique and equipment of the generated output of micro-capacitance sensor photovoltaic generating system | |
| CN110276472A (en) | A kind of offshore wind farm power ultra-short term prediction method based on LSTM deep learning network | |
| CN108846508A (en) | A kind of wind speed forecasting method and system based on atmospheric perturbation | |
| CN112632840A (en) | Power grid transient stability evaluation method based on adaptive differential evolution algorithm and ELM | |
| Fatema et al. | Deterministic and probabilistic occupancy detection with a novel heuristic optimization and Back-Propagation (BP) based algorithm | |
| CN104732067A (en) | Industrial process modeling forecasting method oriented at flow object | |
| CN117786824B (en) | Tunnel environment parameter design method and system based on multi-objective optimization | |
| CN116300755A (en) | Double-layer optimal scheduling method and device for heat storage-containing heating system based on MPC | |
| Zhu et al. | Structural Safety Monitoring of High Arch Dam Using Improved ABC‐BP Model | |
| Ram et al. | Reliability evaluation and cost optimization of solar road studs | |
| CN111027760A (en) | Power load prediction method based on least square vector machine | |
| Xiaojuan et al. | Short-time wind speed prediction for wind farm based on improved neural network | |
| Guo et al. | Predicting the Population Growth and Structure of China Based on Grey Fractional‐Order Models |
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 |
Application publication date: 20181120 |
|
| RJ01 | Rejection of invention patent application after publication |