CN108615097A - A kind of wind speed forecasting method, system, equipment and computer readable storage medium - Google Patents
A kind of wind speed forecasting method, system, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN108615097A CN108615097A CN201810442364.3A CN201810442364A CN108615097A CN 108615097 A CN108615097 A CN 108615097A CN 201810442364 A CN201810442364 A CN 201810442364A CN 108615097 A CN108615097 A CN 108615097A
- Authority
- CN
- China
- Prior art keywords
- speed data
- air speed
- lstm models
- wind speed
- real
- 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
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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/08—Learning methods
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a kind of wind speed forecasting method, system, equipment and computer readable storage medium, wherein this method includes:Obtain real-time air speed data;Real-time air speed data is inputted to LSTM models trained in advance;Obtain the prediction of wind speed data of LSTM models output.A kind of wind speed forecasting method disclosed by the invention, first obtain real-time air speed data, then real-time air speed data is inputted to LSTM models trained in advance, finally obtains the prediction of wind speed data of LSTM models output, namely realizes and wind speed is predicted using LSTM models.It is demonstrated experimentally that the predictablity rate of wind speed forecasting method disclosed by the invention is higher.A kind of forecasting wind speed system, equipment and computer readable storage medium disclosed by the invention also improve the predictablity rate of existing wind speed forecasting method to a certain extent.
Description
Technical field
The present invention relates to nerual network technique fields, more specifically to a kind of wind speed forecasting method, system, equipment
And computer readable storage medium.
Background technology
Wind energy in nature is a kind of reproducible clean energy resource, and the wind energy storage amount in the world is very abundant, this
So that the application of wind energy is more and more wider, in the application process to wind energy, need to predict wind speed.
A kind of existing wind speed forecasting method be based on GRNN (General Regression Neural Network, extensively
Adopted recurrent neural networks) wind speed is predicted.
It is computationally intensive however, existing more to the method Consideration of forecasting wind speed based on GRNN, and predictablity rate compared with
It is low.
In conclusion the predictablity rate for how improving existing wind speed forecasting method be current those skilled in the art urgently
Problem to be solved.
Invention content
The object of the present invention is to provide a kind of wind speed forecasting method, can solve how to improve to a certain extent existing
The technical issues of predictablity rate of wind speed forecasting method.The present invention also provides a kind of forecasting wind speed system, equipment and calculating
Machine readable storage medium storing program for executing.
To achieve the goals above, the present invention provides the following technical solutions:
A kind of wind speed forecasting method, including:
Obtain real-time air speed data;
The real-time air speed data is inputted to LSTM models trained in advance;
Obtain the prediction of wind speed data of the LSTM models output.
Preferably, the LSTM models are trained in advance, including:
Build initial LSTM models;
Obtain training air speed data and test air speed data;
The trained air speed data is inputted to the initial LSTM models;
The test air speed data is input to described by the penalty values that the initial LSTM models are calculated according to loss function
Initial LSTM models obtain accuracy rate;
Whether within a preset range the accuracy rate is judged, if so, the initial LSTM models are determined as instructing in advance
The experienced LSTM models if it is not, then adjusting the weights of the initial LSTM models, and return to the input trained wind speed
The step of data to initial LSTM models.
Preferably, the weights of the adjustment initial LSTM models, including:
The weights of the initial LSTM models are adjusted based on gradient optimization algorithm.
Preferably, the gradient optimization algorithm includes RMSProp optimization algorithms.
Preferably, described to input the real-time air speed data to LSTM models trained in advance, including:
The real-time air speed data is normalized, normalization air speed data is obtained;
The normalization air speed data is inputted to the LSTM models;
The prediction of wind speed data for obtaining the LSTM models output, including:
Obtain the normalization forecasting wind speed data of the LSTM models output;
Renormalization is carried out to the normalization forecasting wind speed data, obtains the prediction of wind speed data.
Preferably, the formula of the normalized includes:
ni=pi/p0-1;
The formula of the renormalization includes:
pi=p0(ni+1);
Wherein, p0Indicate first air speed data of the real-time air speed data;piIndicate the of the real-time air speed data
I air speed data;niThe corresponding normalization air speed data of i-th of data is indicated, in the range of [- 1,1].
Preferably, the real-time air speed data of acquisition, including:
Obtain the real-time air speed data of sensor acquisition.
A kind of forecasting wind speed system, including:
First acquisition module, for obtaining real-time air speed data;
Input module, for inputting the real-time air speed data to LSTM models trained in advance;
Second acquisition module, the prediction of wind speed data for obtaining the LSTM models output.
A kind of forecasting wind speed equipment, including:
Memory, for storing computer program;
Processor, the step of as above any described wind speed forecasting method is realized when for executing the computer program.
A kind of computer readable storage medium is stored with computer program in the computer readable storage medium, described
The step of as above any described wind speed forecasting method is realized when computer program is executed by processor.
A kind of wind speed forecasting method provided by the invention, first obtains real-time air speed data, then inputs real-time air speed data
To LSTM models trained in advance, the prediction of wind speed data of LSTM models output are finally obtained, namely realizes and utilizes LSTM moulds
Type predicts wind speed.It is demonstrated experimentally that the predictablity rate of wind speed forecasting method provided by the invention is higher.The present invention carries
It is pre- that a kind of forecasting wind speed system, equipment and the computer readable storage medium supplied also improves existing wind speed to a certain extent
The predictablity rate of survey method.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of flow chart of wind speed forecasting method provided in an embodiment of the present invention;
Fig. 2 is the flow chart for the weights that initial LSTM models are adjusted based on RMSProp optimization algorithms;
Fig. 3 is a kind of structural schematic diagram of forecasting wind speed system provided in an embodiment of the present invention;
Fig. 4 is a kind of structural schematic diagram of forecasting wind speed equipment provided in an embodiment of the present invention.
Specific implementation mode
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 describes, 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.
The action executive agent of each step can be this hair in a kind of wind speed forecasting method provided in an embodiment of the present invention
A kind of forecasting wind speed system that bright embodiment provides, or built-in computer, server of the forecasting wind speed system etc..
For convenience, here by the action executive agent of each step in a kind of wind speed forecasting method provided in an embodiment of the present invention
It is set as a kind of forecasting wind speed system provided in an embodiment of the present invention, referred to as forecasting system.
Referring to Fig. 1, Fig. 1 is a kind of flow chart of wind speed forecasting method provided in an embodiment of the present invention.
A kind of wind speed forecasting method provided in an embodiment of the present invention, may comprise steps of:
Step S101:Obtain real-time air speed data.
Here real-time air speed data refers to the air speed data that nature generates in real time, the real-time air speed data of acquisition
Quantity can determine that time interval of the real-time air speed data away from prediction of wind speed data can also be according to practical need according to actual needs
It was determined that preferred, real-time air speed data should be in time with the consecutive air speed data of prediction of wind speed data.
Step S102:Real-time air speed data is inputted to LSTM models trained in advance.
Real-time air speed data can be input to LSTM trained in advance by forecasting system after getting real-time air speed data
(Long Short-Term Memory, shot and long term memory network) model, LSTM models can carry out real-time air speed data
Operation obtains prediction of wind speed data corresponding with real-time air speed data.It is the associated description known to the public about LSTM models
The prior art is please referred to, which is not described herein again.
Step S103:Obtain the prediction of wind speed data of LSTM models output.
After LSTM models obtain prediction of wind speed data corresponding with real-time air speed data, forecasting system can obtain
The prediction of wind speed data of LSTM models output.Specifically, can be that LSTM models actively send prediction of wind speed data to predicting and are
System can also be the prediction of wind speed data that forecasting system reads the output of LSTM models in real time, can also be forecasting system with default
Time interval reads the prediction of wind speed data etc. of LSTM models output.It is demonstrated experimentally that a kind of forecasting wind speed side provided by the invention
In method, the mean error of prediction of wind speed data and true air speed data is 0.35395171086315846 (m/s), and prediction is accurate
Rate is high.
A kind of wind speed forecasting method provided by the invention, first obtains real-time air speed data, then inputs real-time air speed data
To LSTM models trained in advance, the prediction of wind speed data of LSTM models output are finally obtained, namely realizes and utilizes LSTM moulds
Type predicts wind speed.It is demonstrated experimentally that the predictablity rate of wind speed forecasting method provided by the invention is higher.
In a kind of wind speed forecasting method provided in an embodiment of the present invention, training LSTM models, are specifically as follows in advance:
Build initial LSTM models;
Obtain training air speed data and test air speed data;
Input trains air speed data to initial LSTM models;
Test air speed data is input to initial LSTM models by the penalty values that initial LSTM models are calculated according to loss function
Obtain accuracy rate;
Judging nicety rate whether within a preset range, if so, initial LSTM models are determined as the LSTM trained in advance
Model if it is not, then adjusting the weights of initial LSTM models, and returns to input training air speed data to the step of initial LSTM models
Suddenly.
In practical application, initial LSTM models, TensorFlow mentioned here can be built by TensorFlow
It is the second generation artificial intelligence learning system that Google is researched and developed based on DistBelief;The initial LSTM models built can be more
It is kind various, for example its input layer can be 50, to hide node layer can be 100, output node layer can be 1,
The activation primitive of hidden layer and output layer can be selected as tanh.Train air speed data and test air speed data that can have multigroup, and
The two group number is identical, and the air speed data quantity of one group of training air speed data can determine according to actual needs, can generally be arranged
Air speed data quantity for the quantity with input layer, one group of test air speed data can be preset quantity, such as 1;
In addition, when obtaining multigroup trained air speed data and test air speed data, sliding microtomy may be used, with one group of training wind speed
The air speed data quantity of data is 50, for the air speed data quantity of one group of test air speed data is 1, is getting a system
After row air speed data, sliding slice can be carried out by being one group in this series of air speed data with 51 data, such as by the 1st
To the 50th data as first group of training air speed data, using the 51st data as first group of test air speed data, by the 2nd
To the 51st data as second group of training air speed data, using the 52nd data as second group of training air speed data, class successively
It pushes away, until obtaining all training air speed datas and test air speed data.In training, one group of training wind can be merely entered every time
Initial LSTM models are trained in fast data to initial LSTM models.When accuracy rate not within a preset range when, adjusting
When the weights of initial LSTM models, exercise wheel number etc. can also be increased, so that accuracy rate can as early as possible within a preset range.
In a kind of wind speed forecasting method provided in an embodiment of the present invention, the weights of initial LSTM models are adjusted, it specifically can be with
For:
The weights of initial LSTM models are adjusted based on gradient optimization algorithm.
In practical application, during adjusting the weights of initial LSTM models, gradient optimization algorithm can be based on
The weights for adjusting initial LSTM models, can improve regulated efficiency to a certain extent in this way.
In a kind of wind speed forecasting method provided in an embodiment of the present invention, gradient optimization algorithm can be specially
RMSProp optimization algorithms.
It is provided the experiment proved that the RMSProp optimization algorithms in gradient optimization algorithm are most suitable for the embodiment of the present invention
A kind of wind speed forecasting method in initial LSTM models weights, using RMSProp optimization algorithms to the power of initial LSTM models
The regulated efficiency highest that value is adjusted, and the predictablity rate of initial LSTM models is best.It is of course also possible to use other ladders
Degree declines optimization algorithm, such as AdaGrad, Adam, to adjust the weights of initial LSTM models.Referring to Fig. 2, being based on
The process that RMSProp optimization algorithms adjust the weights of initial LSTM models may comprise steps of:
Step S201:Determine the initial value of Optimal Parameters θ.
Specifically, the initial value that stochastic variable method of formation generates Optimal Parameters may be used.
Step S202:Determine the learning rate ε and attenuation rate ρ of RMSProp optimization algorithms;Squared gradient r is accumulated in initialization
0。
Step S203:Obtain training air speed data xiAnd estimate air speed data yi。
Here the training air speed data that obtains and air speed data can be that treated data, processing mentioned here is estimated
Including normalization, Gaussian Profile etc..
Step S204:Judge whether to meet stopping criterion, if so, S205 is thened follow the steps, if it is not, thening follow the steps
S206。
Stopping criterion mentioned here can determines according to actual conditions, for example adjustment number accumulation is to a certain threshold value etc..
Step S205:The weights of initial LSTM models are adjusted according to Optimal Parameters.
Step S206:Optimize formula adjusting and optimizing parameter based on RMSProp, returns to step S204, RMSProp optimizations
Formula is:
r←ρr+(l-p)g⊙g;
θ←θ+Δθ;
Wherein, δ is a constant, usually 10-6;M indicates the data bulk of one group of training air speed data;G indicates gradient;r
Indicate accumulation squared gradient;θ indicates Optimal Parameters;xiIndicate i-th group of training air speed data;F indicates xiRelationship between θ;L tables
Show xi、θ、yiBetween relationship;yiIndicate the corresponding test air speed data of i-th group of training air speed data;Local derviation is sought in ▽ expressions;⊙ tables
Show point multiplication operation.
In order to improve the forecasting efficiency of wind speed forecasting method, in a kind of wind speed forecasting method provided in an embodiment of the present invention,
Real-time air speed data is inputted to LSTM models trained in advance, can be specially:
Real-time air speed data is normalized, normalization air speed data is obtained;
Input normalizes air speed data to LSTM models;
The prediction of wind speed data for obtaining the output of LSTM models, are specifically as follows:
Obtain the normalization forecasting wind speed data of LSTM models output;
Renormalization is carried out to normalization forecasting wind speed data, obtains prediction of wind speed data.
Air speed data can be handled by method for normalizing in practical application, the side such as Gaussian Profile can also be used
Method does simplified processing to air speed data, and then improves the forecasting efficiency of wind speed forecasting method.
In a kind of wind speed forecasting method provided in an embodiment of the present invention, the formula of normalized can be:
ni=pi/p0-1;
The formula of renormalization can be:
pi=p0(ni+1);
Wherein, p0Indicate first air speed data of real-time air speed data;piIndicate i-th of wind speed of real-time air speed data
Data;niThe corresponding normalization air speed data of i-th of data is indicated, in the range of [- 1,1].
In practical application, the formula of normalized, the formula of renormalization can be adjusted according to actual needs, this
Invention is not specifically limited herein, for example, can by real-time air speed data maximum value or minimum value etc. be used as p0。
In a kind of wind speed forecasting method provided in an embodiment of the present invention, real-time air speed data is obtained, is specifically as follows:
Obtain the real-time air speed data of sensor acquisition.
In practical application, the real-time air speed data that forecasting system obtains can be the data acquired from sensor.Sensing
Device can with predeterminated frequency, such as 20Hz, sample frequency collect the target wind field period in real-time air speed data.Specifically answer
With in scene, the communication mode between forecasting system and collector can be wired communication mode, or communication
Deng wired communication mode may include:Mobile high definition chained technology (HML), universal serial bus (USB), high-definition multimedia connect
Mouth (HDMI) etc.;Communication may include:Adopting wireless fidelity technology (WiFi), Bluetooth Communication Technology, low-power consumption bluetooth are logical
Letter technology, the communication technology etc. based on IEEE802.11s.
The present invention also provides a kind of forecasting wind speed systems, with a kind of forecasting wind speed side provided in an embodiment of the present invention
The correspondence effect that method has.Referring to Fig. 3, Fig. 3 is a kind of structural representation of forecasting wind speed system provided in an embodiment of the present invention
Figure.
A kind of forecasting wind speed system provided in an embodiment of the present invention may include:
First acquisition module 101, for obtaining real-time air speed data;
Input module 102, for inputting real-time air speed data to LSTM models trained in advance;
Second acquisition module 103, the prediction of wind speed data for obtaining the output of LSTM models.
In a kind of forecasting wind speed system provided in an embodiment of the present invention, including:
Module is built, for building initial LSTM models;
Third acquisition module, for obtaining trained air speed data and test air speed data;
Training data input module, for inputting trained air speed data to initial LSTM models;
Computing module, the penalty values for calculating initial LSTM models according to loss function, by test air speed data input
Accuracy rate is obtained to initial LSTM models;
Judgment module, whether within a preset range for judging nicety rate, if so, initial LSTM models are determined as pre-
First trained LSTM models if it is not, then adjust the weights of initial LSTM models, and return to input training air speed data to initial
The step of LSTM models.
In a kind of forecasting wind speed system provided in an embodiment of the present invention, judgment module may include:
Adjustment unit, the weights for adjusting initial LSTM models based on gradient optimization algorithm.
In a kind of forecasting wind speed system provided in an embodiment of the present invention, gradient optimization algorithm includes RMSProp optimizations
Algorithm.
In a kind of forecasting wind speed system provided in an embodiment of the present invention, input module may include:
Normalization unit obtains normalization air speed data for real-time air speed data to be normalized;
Input unit, for inputting normalization air speed data to LSTM models;
Second acquisition module may include:
Second acquisition unit, the normalization forecasting wind speed data for obtaining the output of LSTM models;
Renormalization unit obtains prediction of wind speed data for carrying out renormalization to normalization forecasting wind speed data.
In a kind of forecasting wind speed system provided in an embodiment of the present invention, the formula of normalized may include:
ni=pi/p0-1;
The formula of renormalization may include:
pi=p0(ni+1);
Wherein, p0Indicate first air speed data of real-time air speed data;piIndicate i-th of wind speed of real-time air speed data
Data;niThe corresponding normalization air speed data of i-th of data is indicated, in the range of [- 1,1].
In a kind of forecasting wind speed system provided in an embodiment of the present invention, the first acquisition module may include:
First acquisition unit, the real-time air speed data for obtaining sensor acquisition.
The present invention also provides a kind of forecasting wind speed equipment and computer readable storage mediums, all have implementation of the present invention
The correspondence effect that a kind of wind speed forecasting method that example provides has.Referring to Fig. 4, Fig. 4 is one kind provided in an embodiment of the present invention
The structural schematic diagram of forecasting wind speed equipment.
A kind of forecasting wind speed equipment provided in an embodiment of the present invention may include:
Memory 201, for storing computer program;
Processor 202 realizes following steps when for executing the computer program preserved in memory 201:
Obtain real-time air speed data;
Real-time air speed data is inputted to LSTM models trained in advance;
Obtain the prediction of wind speed data of LSTM models output.
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:Build initial LSTM models;Obtain training air speed data and test wind speed
Data;Input trains air speed data to initial LSTM models;Obtain initial LSTM models output estimates air speed data;Judge pre-
Estimate air speed data with test air speed data difference whether within a preset range, if so, LSTM models are determined as instructing in advance
Experienced LSTM models if it is not, then adjusting the weights of initial LSTM models, and return to input training air speed data to initial LSTM moulds
The step of type.
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:The weights of initial LSTM models are adjusted based on gradient optimization algorithm.
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:Gradient optimization algorithm includes RMSProp optimization algorithms.
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:Real-time air speed data is normalized, normalization wind speed is obtained
Data;Input normalizes air speed data to LSTM models;Obtain the normalization forecasting wind speed data of LSTM models output;To normalizing
Change forecasting wind speed data and carry out renormalization, obtains prediction of wind speed data.
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:The formula of normalized includes ni=pi/p0-1;The public affairs of renormalization
Formula includes:pi=p0(ni+1);Wherein, p0Indicate first air speed data of real-time air speed data;piIndicate real-time air speed data
I-th of air speed data;niThe corresponding normalization air speed data of i-th of data is indicated, in the range of [- 1,1].
In a kind of forecasting wind speed equipment provided in an embodiment of the present invention, processor 202 executes the meter preserved in memory 201
When calculating loom program, step can be implemented as follows:Obtain the real-time air speed data of sensor acquisition.
A kind of computer readable storage medium provided in an embodiment of the present invention is stored with meter in computer readable storage medium
Calculation machine program, realizes the step of wind speed forecasting method described in any embodiment as above when computer program is executed by processor
Suddenly.
Relevant portion in a kind of forecasting wind speed system provided in an embodiment of the present invention, equipment and computer readable storage medium
Explanation refer to the detailed description of corresponding part in a kind of wind speed forecasting method provided in an embodiment of the present invention, it is no longer superfluous herein
It states.In addition, in above-mentioned technical proposal provided in an embodiment of the present invention with to correspond to technical solution realization principle in the prior art consistent
Part and unspecified, in order to avoid excessively repeat.
The foregoing description of the disclosed embodiments enables those skilled in the art to realize or use the present invention.To this
A variety of modifications of a little embodiments will be apparent for a person skilled in the art, and the general principles defined herein can
Without departing from the spirit or scope of the present invention, to realize in other embodiments.Therefore, the present invention will not be limited
It is formed on the embodiments shown herein, and is to fit to consistent with the principles and novel features disclosed in this article widest
Range.
Claims (10)
1. a kind of wind speed forecasting method, which is characterized in that including:
Obtain real-time air speed data;
The real-time air speed data is inputted to LSTM models trained in advance;
Obtain the prediction of wind speed data of the LSTM models output.
2. according to the method described in claim 1, it is characterized in that, in advance train the LSTM models, including:
Build initial LSTM models;
Obtain training air speed data and test air speed data;
The trained air speed data is inputted to the initial LSTM models;
The test air speed data is input to described initial by the penalty values that the initial LSTM models are calculated according to loss function
LSTM models obtain accuracy rate;
Whether within a preset range to judge the accuracy rate, if so, the initial LSTM models are determined as training in advance
The LSTM models if it is not, then adjusting the weights of the initial LSTM models, and return to the input trained air speed data
The step of to the initial LSTM models.
3. according to the method described in claim 2, it is characterized in that, the weights of the adjustment initial LSTM models, including:
The weights of the initial LSTM models are adjusted based on gradient optimization algorithm.
4. according to the method described in claim 3, it is characterized in that, the gradient optimization algorithm includes RMSProp optimizations
Algorithm.
5. according to the method described in claim 1, it is characterized in that, described input the real-time air speed data to training in advance
LSTM models, including:
The real-time air speed data is normalized, normalization air speed data is obtained;
The normalization air speed data is inputted to the LSTM models;
The prediction of wind speed data for obtaining the LSTM models output, including:
Obtain the normalization forecasting wind speed data of the LSTM models output;
Renormalization is carried out to the normalization forecasting wind speed data, obtains the prediction of wind speed data.
6. according to the method described in claim 2, it is characterized in that, the formula of the normalized includes:
ni=pi/p0-1;
The formula of the renormalization includes:
pi=p0(ni+1);
Wherein, p0Indicate first air speed data of the real-time air speed data;piIndicate i-th of the real-time air speed data
Air speed data;niThe corresponding normalization air speed data of i-th of data is indicated, in the range of [- 1,1].
7. according to the method described in claim 1, it is characterized in that, described obtain real-time air speed data, including:
Obtain the real-time air speed data of sensor acquisition.
8. a kind of forecasting wind speed system, which is characterized in that including:
First acquisition module, for obtaining real-time air speed data;
Input module, for inputting the real-time air speed data to LSTM models trained in advance;
Second acquisition module, the prediction of wind speed data for obtaining the LSTM models output.
9. a kind of forecasting wind speed equipment, which is characterized in that including:
Memory, for storing computer program;
Processor realizes wind speed forecasting method as described in any one of claim 1 to 7 when for executing the computer program
The step of.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer in the computer readable storage medium
Program, the computer program realize wind speed forecasting method as described in any one of claim 1 to 7 when being executed by processor
Step.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810442364.3A CN108615097A (en) | 2018-05-10 | 2018-05-10 | A kind of wind speed forecasting method, system, equipment and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810442364.3A CN108615097A (en) | 2018-05-10 | 2018-05-10 | A kind of wind speed forecasting method, system, equipment and computer readable storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108615097A true CN108615097A (en) | 2018-10-02 |
Family
ID=63662755
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810442364.3A Pending CN108615097A (en) | 2018-05-10 | 2018-05-10 | A kind of wind speed forecasting method, system, equipment and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108615097A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109063939A (en) * | 2018-11-01 | 2018-12-21 | 华中科技大学 | A kind of wind speed forecasting method and system based on neighborhood door shot and long term memory network |
CN110071530A (en) * | 2019-05-20 | 2019-07-30 | 中国矿业大学 | The wind-powered electricity generation of electric system containing energy storage climbing coordinated scheduling method based on LSTM |
CN110210660A (en) * | 2019-05-27 | 2019-09-06 | 河海大学 | A kind of ultra-short term wind speed forecasting method |
CN110309603A (en) * | 2019-07-05 | 2019-10-08 | 华北电力大学(保定) | A kind of short-term wind speed forecasting method and system based on wind speed characteristics |
CN111222677A (en) * | 2019-10-22 | 2020-06-02 | 浙江运达风电股份有限公司 | Wind speed prediction method and system based on long-short term memory time neural network |
CN115796031A (en) * | 2022-11-28 | 2023-03-14 | 中铁四局集团第三建设有限公司 | Tower crane wind speed prediction method, system and computer readable storage medium |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103927460A (en) * | 2014-05-05 | 2014-07-16 | 重庆大学 | Wind power plant short-term wind speed prediction method based on RBF |
-
2018
- 2018-05-10 CN CN201810442364.3A patent/CN108615097A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103927460A (en) * | 2014-05-05 | 2014-07-16 | 重庆大学 | Wind power plant short-term wind speed prediction method based on RBF |
Non-Patent Citations (2)
Title |
---|
刘辉等: ""Smart multi-step deep learning model for wind speed forecasting based on variational mode decomposition, singular spectrum analysis, LSTM network and ELM"", 《ENERGY CONVERSION AND MANAGEMENT DOI: 10.1016/J.ENCONMAN.2018.01.010》 * |
李科: ""基于多元特征融合和LSTM神经网络的中文评论情感分析"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109063939A (en) * | 2018-11-01 | 2018-12-21 | 华中科技大学 | A kind of wind speed forecasting method and system based on neighborhood door shot and long term memory network |
CN110071530A (en) * | 2019-05-20 | 2019-07-30 | 中国矿业大学 | The wind-powered electricity generation of electric system containing energy storage climbing coordinated scheduling method based on LSTM |
CN110210660A (en) * | 2019-05-27 | 2019-09-06 | 河海大学 | A kind of ultra-short term wind speed forecasting method |
CN110210660B (en) * | 2019-05-27 | 2022-07-22 | 河海大学 | Ultra-short-term wind speed prediction method |
CN110309603A (en) * | 2019-07-05 | 2019-10-08 | 华北电力大学(保定) | A kind of short-term wind speed forecasting method and system based on wind speed characteristics |
CN110309603B (en) * | 2019-07-05 | 2023-10-24 | 华北电力大学(保定) | Short-term wind speed prediction method and system based on wind speed characteristics |
CN111222677A (en) * | 2019-10-22 | 2020-06-02 | 浙江运达风电股份有限公司 | Wind speed prediction method and system based on long-short term memory time neural network |
CN115796031A (en) * | 2022-11-28 | 2023-03-14 | 中铁四局集团第三建设有限公司 | Tower crane wind speed prediction method, system and computer readable storage medium |
CN115796031B (en) * | 2022-11-28 | 2023-12-19 | 中铁四局集团第三建设有限公司 | Tower crane wind speed prediction method, system and computer readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108615097A (en) | A kind of wind speed forecasting method, system, equipment and computer readable storage medium | |
US20210383041A1 (en) | In-situ thermodynamic model training | |
Zhang et al. | A comprehensive wind speed prediction system based on Monte Carlo and artificial intelligence algorithms | |
Zhang et al. | Ship motion attitude prediction based on an adaptive dynamic particle swarm optimization algorithm and bidirectional LSTM neural network | |
CN103164742B (en) | A kind of server performance Forecasting Methodology based on particle group optimizing neural network | |
CN111191835B (en) | IES incomplete data load prediction method and system based on C-GAN migration learning | |
CN107704875A (en) | Based on the building load Forecasting Methodology and device for improving IHCMAC neutral nets | |
CN106920006A (en) | A kind of subway station air conditioning energy consumption Forecasting Methodology based on ISOA LSSVM | |
RU2762983C1 (en) | Method and system for monitoring internal air quality and ventilation control for train | |
CN109472397B (en) | Polymerization process parameter adjusting method based on viscosity change | |
Wang et al. | Cell selection with deep reinforcement learning in sparse mobile crowdsensing | |
CN104091216A (en) | Traffic information predication method based on fruit fly optimization least-squares support vector machine | |
CN106779148A (en) | A kind of method for forecasting wind speed of high speed railway line of multi-model multiple features fusion | |
CN109862532B (en) | Rail transit state monitoring multi-sensor node layout optimization method and system | |
CN104075402A (en) | Intelligent air conditioner control method and system | |
CN109598381B (en) | Short-term traffic flow prediction method based on state frequency memory neural network | |
CN105512832A (en) | Urban water demand combined predication method based on time-varying weight minimum variance | |
CN109548044B (en) | DDPG (distributed data group pg) -based bit rate optimization method for energy-collectable communication | |
CN110033081A (en) | A kind of method and apparatus of determining learning rate | |
CN109800884A (en) | Processing method, device, equipment and the computer storage medium of model parameter | |
CN112163671A (en) | New energy scene generation method and system | |
CN111008790A (en) | Hydropower station group power generation electric scheduling rule extraction method | |
CN114322199A (en) | Ventilation system autonomous optimization operation regulation and control platform and method based on digital twins | |
CN113887748B (en) | Online federal learning task allocation method and device, and federal learning method and system | |
CN113821903A (en) | Temperature control method and device, modular data center and storage medium |
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: 20181002 |
|
RJ01 | Rejection of invention patent application after publication |