CN106557840A - A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology - Google Patents
A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology Download PDFInfo
- Publication number
- CN106557840A CN106557840A CN201611029592.5A CN201611029592A CN106557840A CN 106557840 A CN106557840 A CN 106557840A CN 201611029592 A CN201611029592 A CN 201611029592A CN 106557840 A CN106557840 A CN 106557840A
- Authority
- CN
- China
- Prior art keywords
- measuring station
- air measuring
- wind speed
- data
- auxiliary air
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
Abstract
The invention provides a kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, the method is comprised the following steps：Step 1：Auxiliary air measuring station is set；Step 2：Process is filtered to air speed data；Step 3：Filtered each group of data is decomposed；Step 4：Data after decomposing to each group are filtered；Step 5：Each group filtered IMF components and R component are carried out into signal reconstruction；Step 6：Select present period m higher auxiliary air measuring station related to target air measuring station；Step 7：Build the wavelet-neural network model of CS optimizations；Step 8：M auxiliary air measuring station air speed value of actual measurement is input to the model for training, the wind speed value of target air measuring station is obtained.The present invention can carry out high-precision forecast to wind speed along railway under various landform, weather conditions, effectively prevent prediction error and prediction interruption that single air measuring station hardware fault is caused.
Description
Technical field
The invention belongs to railway forecasting wind speed field, more particularly to a kind of high wind line of high-speed railway wind speed adaptive decomposition
Forecasting Methodology.
Background technology
China is vast in territory, seashore line length, and high wind weather often occur in many areas.Wind-powered electricity generation is a kind of valuable resource, so
And, high wind deteriorates can aerodynamic performance of train and lateral stability, the accident such as cause train to be susceptible to derail, topple.High wind
Caused train security incident has generation in countries in the world, to prevent the generation of these accidents, numerous scholars from having carried out items
Research.Wherein, as the dispatch control of railway interests must possess accuracy and advanced, therefore, forecasting wind speed along railway
Become one of research contents of core.
Wind speed is affected by various factors, with very strong randomness and non-stationary, is predicted very difficult.Traditional wind
Fast predictive study focuses mostly in predicting wind speed of wind farm, method mainly have time serieses, Kalman filtering, artificial neural network,
Support vector machine, wavelet analysises, empirical mode decomposition etc., in recent years, some scholars propose new built-up pattern, these combinations
The estimated performance of model is better than single model.
Because under high wind conditions, compare other sections, train by some particular sections (bridge such as air port region,
High embankment, hills, curve etc.) when easily there is security incident, so, to these particular sections, forecasting wind speed is more accurate
Really.At present, no matter forecasting wind speed along railway method particular section or common section, are mostly based on single air measuring station wind speed number
According to sampled signal is single, and anti-interference is poor, it is impossible to avoid prediction error that single air measuring station hardware fault causes and prediction from interrupting.
The existing wind speed forecasting method based on spatial coherence more be directed to large-scale wind power field, and or each air measuring station position be separated by compared with
Far, precision of prediction is not high；Or analyzed using big time scale, short-term forecast precision is not high；Or flow field analysis method is adopted,
The calculating time is long；It is difficult to fully meet the requirement of railway dispatching commander.Therefore, in the urgent need to set up one kind can meet zonule,
The wind speed along railway wisdom Forecasting Methodology that short-term, high accuracy needs.
The content of the invention
The invention provides a kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, it is intended that gram
Take not enough present in existing railway wind speed forecasting method, precision of prediction is improved by many air measuring station data, simultaneously because fusion
Air speed data multiformity between multiple air measuring stations, it is ensured that the stability of forecast model, and it is avoided that single air measuring station sensor is hard
The prediction that part failure is caused is interrupted.
A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, comprises the following steps：
Step 1：N number of auxiliary air measuring station is at least installed around target air measuring station position, is adopted using auxiliary air measuring station in real time
The air speed data of collection target air measuring station, obtains the wind speed sample set of target air measuring station and auxiliary air measuring station；
Wherein, N is the integer more than or equal to 5；
Auxiliary air measuring station, is searched for 20 kilometers as maximum radius along Along Railway with target air measuring station position as origin
Railway contact line column around before and after target air measuring station position；By it is N number of auxiliary air measuring station be arranged on railway contact line column away from
On the height and position that 4 meters of railway rail level；Target air measuring station and auxiliary air measuring station position height and roof of traveling train etc. along the line
It is high；
Step 2：Process is filtered to the air speed data in wind speed sample set and removes potential error in air speed data；
Step 3：To being entered using adaptive noise completely integrated empirical modal through the filtered wind speed sample set of step 2
Row decomposes；
Step 4：Process is filtered to the data Jing after step 3 is decomposed；
Step 5：Jing steps 4 filtered data are carried out into signal reconstruction, the wind speed reconstruct data of each air measuring station are obtained；
Step 6：The wind speed reconstruct data of each auxiliary air measuring station and the wind speed of target air measuring station reconstruct data are carried out related
Property inspection, sorted by degree of association from high to low, select and target air measuring station wind speed reconstruct Data mutuality degree ranking before m groups auxiliary survey
The wind speed reconstruct data at wind station and corresponding m auxiliary air measuring station；
Wherein, m is integer, and span is [3,60%N]；
Step 7：By target air measuring station and step 6 select auxiliary air measuring station wind speed reconstruct data according to frequency partition into
High frequency subsequence, intermediate frequency subsequence, low frequency subsequence, then will be the high frequency subsequence of all air measuring stations, intermediate frequency subsequence, low frequency
Sequence is referred to high frequency layer, intermediate frequency layer and low frequency layer respectively；
Step 8：The prediction model based on wavelet neural network based on CS of each frequency layer is trained, by the output of each frequency layer forecast model
As a result the output result added up as overall forecast model, obtains overall forecast model；
Using in high frequency layer, the high frequency subsequence of each auxiliary air measuring station and the high frequency subsequence of target air measuring station are as height
Input and output of the frequency layer based on the prediction model based on wavelet neural network of CS；
Using in intermediate frequency layer, the intermediate frequency subsequence of each auxiliary air measuring station and the intermediate frequency subsequence of target air measuring station are as in
Input and output of the frequency layer based on the prediction model based on wavelet neural network of CS；
Using in low frequency layer, each low frequency subsequence of auxiliary air measuring station and the low frequency subsequence of target air measuring station are as low
Input and output of the frequency layer based on the prediction model based on wavelet neural network of CS；
Step 9：By the air speed data of the auxiliary air measuring station of Real-time Collection, after being processed according to step 2-5, and to processing
Data afterwards carry out frequency partition according to step 7, in the overall forecast model that the data input after division to step 8 is built,
Obtain the forecasting wind speed result of target air measuring station.
The filtering method adopted in the step 2 and step 4 is Interactive Multiple-Model Kalman filtering method.
Correlation test method used in the step 6 goes trend cross-correlation analysis method for DCCA.
The construction step of the prediction model based on wavelet neural network based on CS is as follows：
Step 1：Initial network connection weight and the bird of threshold value of 50 groups of wavelet neural networks are randomly generated using CS algorithms
Nest position population；
Wherein, each Bird's Nest position correspond to the initial connection weight and threshold value of one group of wavelet neural network, Bird's Nest position
The renewal algebraically put is 100；
Step 2：According to forecasting wind speed root-mean-square error minimum principle, it is right each Bird's Nest position to be carried out using wind speed sample
Than filtering out the Bird's Nest position of optimum；
In screening every time, the 10 groups of birds for randomly generating again are used to 10 groups of Bird's Nests most bad in original 50 groups of Bird's Nest positions
Nest is changed, and 5 groups of Bird's Nests optimum in 50 groups of Bird's Nest positions of previous step are carried out remaining into the performance comparison of next step
In；
Step 3：After 100 step is reached, CS algorithms export optimal Bird's Nest position, by the optimum of this step initial connection weight
Wavelet-neural network model is given to threshold value.
Step 4：The input of each subsequence of air measuring station as wavelet-neural network model, correspondence target air measuring station will be aided in
Each subsequence as output, complete the learning training process of itself, obtain the prediction model based on wavelet neural network based on CS；
Wherein, based on CS prediction model based on wavelet neural network hidden layer node transmission function be wavelet mother function.
It is to produce wavelet neural network optimum using the purpose of CS algorithms (i.e. cuckoo algorithm) Optimization of Wavelet neutral net
Initial network connection weight and hidden layer threshold value.
Beneficial effect
The invention discloses a kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, the method includes following
Step：Step 1：Auxiliary air measuring station is set；Step 2：Process is filtered to air speed data；Step 3：By filtered each group number
According to being decomposed；Step 4：Data after decomposing to each group are filtered；Step 5：By each group filtered IMF components and R point
Amount carries out signal reconstruction；Step 6：Select present period m higher auxiliary air measuring station related to target air measuring station；Step 7：Structure
Build the wavelet-neural network model of CS optimizations；Step 8：M auxiliary air measuring station air speed value of actual measurement is input to the mould for training
Type, obtains the wind speed value of target air measuring station.The present invention can be carried out to wind speed along railway under various landform, weather conditions
High-precision forecast, effectively prevent prediction error and prediction interruption that single air measuring station hardware fault is caused.
By 5 auxiliary air measuring stations are installed around target air measuring station position, by aiding in air measuring station Real-time Collection per second
Wind field air speed data around target air measuring station, forms including the wind speed sample set including target air measuring station and auxiliary air measuring station
Close；To the Interactive Multiple-Model Kalman filtering process of original air speed data, the measurement error that sensor brings is rejected, is utilized
CEEMDAN decomposes, and rejects signal high frequency Jump；Spend trend cross-correlation analysis method present period is accurately selected with target survey
The higher auxiliary air measuring station of wind station wind velocity signal degree of association；The wavelet neural network optimized using CS is trained and is predicted.
The ingenious part of patent of the present invention is：First, different from itself being obtained to target air measuring station in prior art
The wind speed time serieses for obtaining directly set up forecast model, but dexterously target are surveyed using the wavelet neural network after CS optimizations
The auxiliary air measuring station of wind station and its periphery sets up non-linear forecasting wind speed network, is sufficiently used and is gathered along the railway
Space air speed data；Secondly, in order to reduce potential precision of prediction of the measurement error to wavelet neural network in air speed data
Affect, dexterously propose the technical scheme for decomposing mixed processing with the filtering of Interactive Multiple-Model karr Man U and CEEMDAN, and
Be not it is conventional be directly filtered process, i.e., enter first with whole original air speed data more than Interactive Multiple-Model Kalman filtering
Row removes the Filtering Processing of noise, then carries out CEEMDAN decomposition to filtered whole air speed datas for the first time, finally in fortune
Second Filtering Processing is carried out to the high frequency air speed data after decomposition with Interactive Multiple-Model Kalman filtering, by filtering, decomposing,
The process for filtering again, it is ensured that for setting up the pure property of the air speed data of wavelet neural network；3rd, this patent is without simple
Ground is directly inputted to wavelet-neural network model the air speed data after decomposition, but they are divided into according to own frequency characteristic
Three classes, can not only so improve the learning performance of wavelet-neural network model, and improve the output of forecast model
Real-time.
Description of the drawings
Principle flow charts of the Fig. 1 for the method for the invention；
Structure charts of the Fig. 2 for the prediction network inputs output wind speed data of the method for the invention；
Fig. 3 is by the railway forecasting wind speed result that obtained using Forecasting Methodology proposed by the invention；
Fig. 4 is by the railway forecasting wind speed result that obtained using traditional single MLP neural network model；
Fig. 5 is the railway forecasting wind speed result obtained using traditional single ARIMA model.
Specific embodiment
Below in conjunction with drawings and Examples, the present invention is described further.
As shown in figure 1, a kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, comprises the following steps：
Step 1：It is that the railway future wind speed to certain target air measuring station position realizes prediction, pacifies around the air measuring station position
Fill 5 auxiliary air measuring stations.Obtain the original air speed data of same session target air measuring station and 5 auxiliary air measuring stations, every group of wind speed
Packet contains 600 data, and first 500 in 600 data are used to model, and the 501st～600 data are used to verify.
Target air measuring station is designated as into A, 5 auxiliary air measuring stations are designated as B, C, D, E, F respectively, before each air measuring station 500 it is original
Air speed data is expressed as follows：
The original air speed data of target air measuring station A：{a1,a2,a3...,a499,a500}
The original air speed data of auxiliary air measuring station B：{b1,b2,b3...,b499,b500}
The original air speed data of auxiliary air measuring station C：{c1,c2,c3...,c499,c500}
The original air speed data of auxiliary air measuring station D：{d1,d2,d3...,d499,d500}
The original air speed data of auxiliary air measuring station E：{e1,e2,e3...,e499,e500}
The original air speed data of auxiliary air measuring station F：{f1,f2,f3...,f499,f500}
Step 2：With Interactive Multiple-Model Kalman filtering method to air measuring station A, the original air speed data of B, C, D, E, F is filtered
Ripple process, removes potential error in air speed data, obtains following filtered air speed datas：
The filtered air speed datas of target air measuring station A：{a′1,a′2,a′3...,a′499,a′500}
The filtered air speed datas of auxiliary air measuring station B：{b′1,b′2,b′3...,b′499,b′500}
The filtered air speed datas of auxiliary air measuring station C：{c′1,c′2,c′3...,c′499,c′500}
The filtered air speed datas of auxiliary air measuring station D：{d′1,d′2,d′3...,d′499,d′500}
The filtered air speed datas of auxiliary air measuring station E：{e′1,e′2,e′3...,e′499,e′500}
The filtered air speed datas of auxiliary air measuring station F：{f′1,f′2,f′3...,f′499,f′500}
Step 3：Using adaptive noise completely integrated empirical mode decomposition, reduce reconstructed error, obtain following components：
Target air measuring station A：A′IMF1,A′IMF2,...A′IMF8,A′R
Auxiliary air measuring station B：B′IMF1,B′IMF2,...B′IMF8,B′R
Auxiliary air measuring station C：C′IMF1,C′IMF2,...C′IMFn,C′R
Auxiliary air measuring station D：D′IMF1,D′IMF2,...D′IMF8,D′R
Auxiliary air measuring station E：E′IMF1,E′IMF2,...E′IMF8,E′R
Auxiliary air measuring station F：F′IMF1,F′IMF2,...F′IMF8,F′R
Step 4：The IMF components comprising high frequency Jump obtained to each component solution continue using Interactive Multiple-Model karr
Graceful filter method is filtered process, to remove the high frequency Jump in IMF components, obtains following components：
Target air measuring station A：AIMF1,AIMF2,...AIMF8,AR
Auxiliary air measuring station B：BIMF1,BIMF2,...BIMF8,BR
Auxiliary air measuring station C：CIMF1,CIMF2,...CIMF8,CR
Auxiliary air measuring station D：DIMF1,DIMF2,...DIMF8,DR
Auxiliary air measuring station E：EIMF1,EIMF2,...EIMF8,ER
Auxiliary air measuring station F：FIMF1,FIMF2,...FIMF8,FR
Step 5：Filtered IMF components and R component are carried out into signal reconstruction, the wind speed number after following signal reconstructions is obtained
According to：
Target air measuring station A：{a″1,a″2,a″3...,a″499,a″500}
Auxiliary air measuring station B：{b″1,b″2,b″3...,b″499,b″500}
Auxiliary air measuring station C：{c″1,c″2,c″3...,c″499,c″500}
Auxiliary air measuring station D：{d″1,d″2,d″3...,d″499,d″500}
Auxiliary air measuring station E：{e″1,e″2,e″3...,e″499,e″500}
Auxiliary air measuring station F：{f″1,f″2,f″3...,f″499,f″500}
Step 6：By it is each auxiliary air measuring station signal reconstruction after air speed data by go trend cross-correlation analysis method respectively and
Air speed data after target air measuring station signal reconstruction carries out correlation test, and 5 groups of data of auxiliary air measuring station are carried out by group
Relevancy ranking, selects the maximum front 3 groups of data of degree of association and its corresponding 3 auxiliary air measuring station.
This 3 auxiliary air measuring stations selected in this example respectively aid in air measuring station B, auxiliary air measuring station C, auxiliary air measuring station
D。
Step 7：As shown in Fig. 2 to the target air measuring station A and filtered IMF of each group of auxiliary air measuring station B, C, D for selecting
Component and R component are divided into high frequency subsequence IMF by frequency1,IMF2,IMF3, intermediate frequency subsequence IMF4,IMF5,IMF6, the sub- sequence of low frequency
Row IMF7,IMF8,IMFR.Each group subsequence is layered according to frequency, be finally obtained 1 high frequency layer, 1 intermediate frequency layer, 1
Individual low frequency layer.CS (cuckoo searching algorithm) optimizations are set up respectively to 1 high frequency layer obtaining, 1 intermediate frequency layer, 1 low frequency layer
Wavelet neural network.
High frequency layer is with the high frequency subsequence IMF of 3 auxiliary air measuring stations1、IMF2、IMF3To be input into, with the height of target air measuring station
Frequency subsequence IMF1、IMF2、IMF3For output；
Intermediate frequency layer is with the intermediate frequency subsequence IMF of 3 auxiliary air measuring stations4、IMF5、IMF6To be input into, with target air measuring station
Frequency subsequence IMF4、IMF5、IMF6For output；
Low frequency layer is with the low frequency subsequence IMF of 3 auxiliary air measuring stations7、IMF8、IMFRTo be input into, with the low of target air measuring station
Frequency subsequence IMF7、IMF8、IMFRFor output；Network is trained.
Step 8：After model training good, the air speed data of actual measurement B, C, D auxiliary air measuring station for checking is entered successively
The above-mentioned Interactive Multiple-Model Kalman filtering of row, the complete empirical mode decomposition of adaptive noise, to IMF1,IMF2,IMF3Component is carried out
Interactive Multiple-Model Kalman filtering, then the wavelet neural that the data input for obtaining is optimized to the cuckoo searching algorithm for training
Network, predicts the IMF components of target air measuring station A, then collects the wind speed value for being calculated target air measuring station A
By target air measuring station A wind speed valuesWith target air measuring station A wind speed measured datas
{a501,a502,a503...,a599,a600Contrasted, the prediction effect of testing model.
Realize that using Forecasting Methodology proposed by the invention the result of forecasting wind speed is as shown in Figure 3.Using traditional single MLP
Neural network model realizes that the result of forecasting wind speed is as shown in Figure 4.Using the knot of traditional single ARIMA model realization forecasting wind speed
Fruit is as shown in Figure 5.Precision index calculating is carried out to predicting the outcome shown in Fig. 3-Fig. 5 using formula (1-3), 1 He is the results are shown in Table
Table 2.
Mean absolute error:
Mean absolute relative error:
Root-mean-square error:
In above-mentioned formula, n is the air speed data number for model testing, and it is 100 that this patent takes n.X (i) is actual measurement wind
Fast data,For prediction of wind speed data.
Table 1：Using the precision of prediction of the proposed Forecasting Methodology of this patent
| Mean absolute error | 0.4129m/s |
| Mean absolute relative error | 2.53% |
| Root-mean-square error | 0.5912m/s |
Table 2：Using the precision of prediction of traditional MLP neural network models
| Mean absolute error | 1.4102m/s |
| Mean absolute relative error | 8.80% |
| Root-mean-square error | 2.0551m/s |
Table 3：Using the precision of prediction of traditional ARIMA models
| Mean absolute error | 1.3160m/s |
| Mean absolute relative error | 8.60% |
| Root-mean-square error | 1.7486m/s |
From Fig. 3, Fig. 4 and Fig. 5, and with reference to Tables 1 and 2 from the point of view of, method of the present invention, from mean absolute error, flat
From the point of view of equal absolute relative error and root-mean-square error, hence it is evident that better than prior art, show that the method for the invention has preferable
Application effect.
Claims (4)
1. a kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology, it is characterised in that comprise the following steps：
Step 1：N number of auxiliary air measuring station is at least installed around target air measuring station position, using auxiliary air measuring station Real-time Collection mesh
The air speed data of mark air measuring station, obtains the wind speed sample set of target air measuring station and auxiliary air measuring station；
Wherein, N is the integer more than or equal to 5；
Step 2：Process is filtered to the air speed data in wind speed sample set and removes potential error in air speed data；
Step 3：To being carried out point using adaptive noise completely integrated empirical modal through the filtered wind speed sample set of step 2
Solution；
Step 4：Process is filtered to the data Jing after step 3 is decomposed；
Step 5：Jing steps 4 filtered data are carried out into signal reconstruction, the wind speed reconstruct data of each air measuring station are obtained；
Step 6：Wind speed reconstruct data of the wind speed reconstruct data of each auxiliary air measuring station with target air measuring station are carried out into dependency inspection
Test, sorted by degree of association from high to low, select and m groups auxiliary air measuring station before target air measuring station wind speed reconstruct Data mutuality degree ranking
Wind speed reconstruct data and it is corresponding m aid in air measuring station；
Wherein, m is integer, and span is [3,60%N]；
Step 7：The wind speed of the auxiliary air measuring station that target air measuring station and step 6 are selected reconstructs data according to frequency partition into high frequency
Subsequence, intermediate frequency subsequence, low frequency subsequence, then by the high frequency subsequence of all air measuring stations, intermediate frequency subsequence, low frequency subsequence
High frequency layer, intermediate frequency layer and low frequency layer are referred to respectively；
Step 8：The prediction model based on wavelet neural network based on CS of each frequency layer is trained, by the output result of each frequency layer forecast model
The cumulative output result as overall forecast model, obtains overall forecast model；
Using in high frequency layer, the high frequency subsequence of each auxiliary air measuring station and the high frequency subsequence of target air measuring station are as high frequency layer
Input and output based on the prediction model based on wavelet neural network of CS；
Using in intermediate frequency layer, the intermediate frequency subsequence of each auxiliary air measuring station and the intermediate frequency subsequence of target air measuring station are as intermediate frequency layer
Input and output based on the prediction model based on wavelet neural network of CS；
Using in low frequency layer, the low frequency subsequence of each auxiliary air measuring station and the low frequency subsequence of target air measuring station are as low frequency layer
Input and output based on the prediction model based on wavelet neural network of CS；
Step 9：By Real-time Collection auxiliary air measuring station air speed data, after being processed according to step 2-5, and to process after
Data carry out frequency partition according to step 7, in the overall forecast model that the data input after division to step 8 is built, obtain
The forecasting wind speed result of target air measuring station.
2. method according to claim 1, it is characterised in that the filtering method adopted in the step 2 and step 4 is to hand over
Mutual multi-model Kalman filtering method.
3. method according to claim 1 and 2, it is characterised in that the correlation test method used in the step 6 is
DCCA goes trend cross-correlation analysis method.
4. method according to claim 3, it is characterised in that the structure of the prediction model based on wavelet neural network based on CS
Build step as follows：
Step 1：The Bird's Nest position of the initial network connection weight and threshold value of 50 groups of wavelet neural networks is randomly generated using CS algorithms
Put population；
Wherein, each Bird's Nest position correspond to the initial connection weight and threshold value of one group of wavelet neural network, Bird's Nest position
It is 100 to update algebraically；
Step 2：According to forecasting wind speed root-mean-square error minimum principle, each Bird's Nest position is contrasted using wind speed sample,
Filter out the Bird's Nest position of optimum；
In screening every time, 10 groups of Bird's Nests most bad in original 50 groups of Bird's Nest positions are entered with the 10 groups of Bird's Nests for randomly generating again
Row is changed, and 5 groups of Bird's Nests optimum in 50 groups of Bird's Nest positions of previous step are carried out in the performance comparison for remain into next step；
Step 3：After 100 step is reached, CS algorithms export optimal Bird's Nest position, by the optimum of this step initial connection weight and threshold
Value is given to wavelet-neural network model；
Step 4：The input of each subsequence of air measuring station as wavelet-neural network model will be aided in, each of target air measuring station will be corresponded to
Subsequence completes the learning training process of itself, obtains the prediction model based on wavelet neural network based on CS as output；
Wherein, based on CS prediction model based on wavelet neural network hidden layer node transmission function be wavelet mother function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611029592.5A CN106557840B (en) | 2016-11-14 | 2016-11-14 | A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611029592.5A CN106557840B (en) | 2016-11-14 | 2016-11-14 | A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106557840A true CN106557840A (en) | 2017-04-05 |
| CN106557840B CN106557840B (en) | 2018-01-09 |
Family
ID=58444336
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611029592.5A Active CN106557840B (en) | 2016-11-14 | 2016-11-14 | A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106557840B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108985520A (en) * | 2018-08-02 | 2018-12-11 | 广东工业大学 | A kind of wind speed forecasting method, device and equipment |
| CN109190839A (en) * | 2018-09-20 | 2019-01-11 | 中南大学 | A kind of Along Railway wind speed Intelligent rolling prediction technique merging wind direction |
| CN109299430A (en) * | 2018-09-30 | 2019-02-01 | 淮阴工学院 | The short-term wind speed forecasting method with extreme learning machine is decomposed based on two stages |
| CN109709934A (en) * | 2018-12-11 | 2019-05-03 | 南京航空航天大学 | A kind of flight control system fault diagnosis redundancy design method |
| CN110475043A (en) * | 2019-07-31 | 2019-11-19 | 西安工程大学 | A kind of conversion method of CMYK to Lab color space |
| CN113191524A (en) * | 2021-01-29 | 2021-07-30 | 河北建投新能源有限公司 | Wind speed prediction method and device, storage medium and processor |
| CN113325472A (en) * | 2021-05-21 | 2021-08-31 | 成都理工大学 | Seismic wave field sub-component extraction method based on principal component analysis |
| CN113688770A (en) * | 2021-09-02 | 2021-11-23 | 重庆大学 | Long-term wind pressure missing data completion method and device for high-rise building |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542133A (en) * | 2010-12-10 | 2012-07-04 | 中国科学院深圳先进技术研究院 | Short-time wind speed forecasting method and system for wind power plant |
| JP2014128184A (en) * | 2012-12-27 | 2014-07-07 | Mitsubishi Heavy Ind Ltd | Control device and control method for wind generator system, and wind generator system |
-
2016
- 2016-11-14 CN CN201611029592.5A patent/CN106557840B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102542133A (en) * | 2010-12-10 | 2012-07-04 | 中国科学院深圳先进技术研究院 | Short-time wind speed forecasting method and system for wind power plant |
| JP2014128184A (en) * | 2012-12-27 | 2014-07-07 | Mitsubishi Heavy Ind Ltd | Control device and control method for wind generator system, and wind generator system |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108985520A (en) * | 2018-08-02 | 2018-12-11 | 广东工业大学 | A kind of wind speed forecasting method, device and equipment |
| CN109190839A (en) * | 2018-09-20 | 2019-01-11 | 中南大学 | A kind of Along Railway wind speed Intelligent rolling prediction technique merging wind direction |
| CN109299430A (en) * | 2018-09-30 | 2019-02-01 | 淮阴工学院 | The short-term wind speed forecasting method with extreme learning machine is decomposed based on two stages |
| CN109709934A (en) * | 2018-12-11 | 2019-05-03 | 南京航空航天大学 | A kind of flight control system fault diagnosis redundancy design method |
| CN109709934B (en) * | 2018-12-11 | 2021-04-06 | 南京航空航天大学 | Fault diagnosis redundancy design method for flight control system |
| CN110475043A (en) * | 2019-07-31 | 2019-11-19 | 西安工程大学 | A kind of conversion method of CMYK to Lab color space |
| CN110475043B (en) * | 2019-07-31 | 2021-08-17 | 西安工程大学 | Method for converting CMYK to Lab color space |
| CN113191524A (en) * | 2021-01-29 | 2021-07-30 | 河北建投新能源有限公司 | Wind speed prediction method and device, storage medium and processor |
| CN113191524B (en) * | 2021-01-29 | 2023-07-11 | 河北建投新能源有限公司 | Wind speed prediction method, device, storage medium and processor |
| CN113325472A (en) * | 2021-05-21 | 2021-08-31 | 成都理工大学 | Seismic wave field sub-component extraction method based on principal component analysis |
| CN113688770A (en) * | 2021-09-02 | 2021-11-23 | 重庆大学 | Long-term wind pressure missing data completion method and device for high-rise building |
| CN113688770B (en) * | 2021-09-02 | 2024-02-06 | 重庆大学 | Method and device for supplementing long-term wind pressure missing data of high-rise building |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106557840B (en) | 2018-01-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106557840B (en) | A kind of high wind line of high-speed railway wind speed adaptive decomposition Forecasting Methodology | |
| CN106779151B (en) | A kind of line of high-speed railway wind speed multi-point multi-layer coupling prediction method | |
| CN106372731B (en) | A kind of high wind line of high-speed railway wind speed spatial network structure forecast method | |
| CN106772695B (en) | A kind of Along Railway wind speed forecasting method for merging many air measuring station measured datas | |
| CN106779148B (en) | A kind of method for forecasting wind speed of high speed railway line of multi-model multiple features fusion | |
| Zhao et al. | Multi-step wind speed and power forecasts based on a WRF simulation and an optimized association method | |
| CN103324980B (en) | A kind of method for forecasting | |
| CN105184391B (en) | Wind farm wind velocity and power forecasting method based on wavelet decomposition and support vector machines | |
| CN102129511B (en) | System for forecasting short-term wind speed of wind power station based on MATLAB | |
| Zhang et al. | Short term wind energy prediction model based on data decomposition and optimized LSSVM | |
| CN107704953A (en) | The short-term wind-electricity power probability density Forecasting Methodology of EWT quantile estimate forests | |
| CN102411729B (en) | Wind power prediction method based on adaptive linear logic network | |
| CN107169598A (en) | A kind of weather prediction method and system based on deep learning | |
| CN109948833A (en) | A kind of Hydropower Unit degradation trend prediction technique based on shot and long term memory network | |
| CN103577694A (en) | Aquaculture water quality short-time combination forecast method on basis of multi-scale analysis | |
| CN107480698A (en) | Method of quality control based on multiple monitoring indexes | |
| CN106920007A (en) | PM based on second order Self-organized Fuzzy Neural Network2.5Intelligent Forecasting | |
| CN102830250B (en) | Method for diagnosing faults of wind speed sensor at wind power plant based on spatial relevancy | |
| CN105205495A (en) | Non-stationary fluctuating wind speed forecasting method based on EMD-ELM | |
| CN103020743B (en) | Wind energy turbine set ultra-short term wind speed forecasting method | |
| CN106229972A (en) | A kind of integrated based on many meteorological sources and the wind power forecasting method of segmentation modeling | |
| CN107563566A (en) | A kind of run time interval prediction method between bus station based on SVMs | |
| CN107895202A (en) | Short-term wind speed forecasting method based on various visual angles wind speed mode excavation | |
| CN106897794A (en) | A kind of wind speed forecasting method based on complete overall experience mode decomposition and extreme learning machine | |
| CN107180534A (en) | The express highway section average speed method of estimation of support vector regression fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |