CN105205495A - Non-stationary fluctuating wind speed forecasting method based on EMD-ELM - Google Patents

Non-stationary fluctuating wind speed forecasting method based on EMD-ELM Download PDF

Info

Publication number
CN105205495A
CN105205495A CN201510556369.5A CN201510556369A CN105205495A CN 105205495 A CN105205495 A CN 105205495A CN 201510556369 A CN201510556369 A CN 201510556369A CN 105205495 A CN105205495 A CN 105205495A
Authority
CN
China
Prior art keywords
wind speed
stationary
fluctuating wind
stationary fluctuating
emd
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
Application number
CN201510556369.5A
Other languages
Chinese (zh)
Inventor
李春祥
钟旺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Shanghai for Science and Technology
Original Assignee
University of Shanghai for Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Shanghai for Science and Technology filed Critical University of Shanghai for Science and Technology
Priority to CN201510556369.5A priority Critical patent/CN105205495A/en
Publication of CN105205495A publication Critical patent/CN105205495A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines

Abstract

The invention provides a non-stationary fluctuating wind speed forecasting method based on EMD-ELM. The method comprises the following steps: generating a non-stationary fluctuating wind speed sample by utilizing a TARMA model in a simulated manner, and dividing the fluctuating wind speed sample into a training set and a test set two parts, and carrying out normalization processing on the sample by utilizing Matlab; carrying out EMD processing on time sequence of the non-stationary fluctuating wind speed sample, and decomposing non-stationary nonlinear fluctuating wind signals into a group of stationary and linear sequence sets; carrying out phase-space reconstruction on the group of IMFs, and establishing corresponding ELM prediction models respectively; and carrying out superposition on the forecast results of the group of IMFs to obtain forecasted non-stationary fluctuating wind speed of the point, meanwhile, comparing the test sample with the forecasted non-stationary fluctuating wind speed result, and calculating average error, root-mean-square error and correlation coefficient of the forecasted wind speed and the actual wind speed. The invention provides a fast-speed and good-effect method for non-stationary fluctuating wind speed prediction.

Description

Based on the non-stationary fluctuating wind speed Forecasting Methodology of EMD-ELM
Technical field
The present invention relates to a kind of single-point non-stationary fluctuating wind speed Forecasting Methodology adopting empirical mode decomposition (EMD) and extreme learning machine (ELM) to combine, a kind of non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM specifically.
Background technology
For large span spatial structure, Longspan Bridge, high building structure, tall and slender structure, as buildingss such as stay-supported mast, television tower, chimneys, wind load is one of control load of structural wind resistance design.And first the Wind resistant analysis of carrying out structure will obtain the sample data of wind load, determine that the main research means of Wind Engineering has theoretical analysis, numerical simulation, wind tunnel test and field measurement etc. at present.Along with the develop rapidly of computer technology and people are to the further investigation of stochastic process numerical simulation technology, adopt method for numerical simulation the to obtain arbitrariness of the condition such as feature that Wind Velocity History curve can consider place, wind spectrum signature, buildings, the load that simulation is obtained is as far as possible close to the actual wind-force of structure, the arbitrariness of some statistical property can be met simultaneously, and more representative than physical record, be thus widely used in Practical Project.
Non-stationary property as the ubiquitous a kind of phenomenon of the various random load of occurring in nature, as atmospheric turbulence of boundary layer, thunderstorm high wind and earthquake etc.Its amplitude and frequency are all time dependent, therefore under some specific environment, carry out numerical simulation to fluctuating wind, and the non-stationary of wind is the factor that must consider.Particularly in downburst, namely strong in Thunderstorm Weather down draft clashes ground, and propagated along earth's surface to surrounding by rum point have sudden and destructive a kind of high wind, it is extremely strong non-stationaryly produces larger dynamic response to structure possibly.A large amount of actual test data analysis shows, under the complicated landform of harsher wind conditions, many wind speed records do not meet this stationarity requirement.Non-stationary fluctuating wind particularly under complicated landform harsher wind conditions, when adopting steady wind speed to suppose, Non-stationary Data needs to give up, and this can cause larger analytical error, as turbulence intensity value can be over-evaluated, and then affects the accuracy of subsequent analysis.
Extreme learning machine (ExtremeLearningMachine, ELM) be a kind of novel Single hidden layer feedforward neural networks (SLFNs) learning method, this algorithm is under the prerequisite of Stochastic choice input layer weights and hidden neuron threshold value, calculate by means of only a step and can try to achieve network output weights, compare with traditional neural network, extreme learning machine drastically increases generalization ability and the pace of learning of network, has stronger nonlinear fitting ability.Its basic thought is: arrange suitable the number of hidden nodes before training, and only need for input weights and hidden layer are biased random assignment in the process of implementation, whole process once completes, and without the need to iteration, and produces unique optimum solution.Fluctuating wind speed simulation generated, as learning training sample, is set up regression model and is effectively predicted single-point fluctuating wind speed.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM, autoregressive moving-average model (Time-VaryingAuto-RegressiveandMovingAverage is become when it utilizes, TARMA) simulation generates non-stationary fluctuating wind speed sample, based on empirical mode decomposition (EMD) and sort reorganization method, set up the model of extreme learning machine (ELM), utilize this model to predict single-point non-stationary fluctuating wind speed.Calculate the validity that the average error (AE) of actual wind speed and prediction of wind speed, root-mean-square error (RMSE) and related coefficient (R) evaluate this method simultaneously.
According to foregoing invention design, the present invention adopts following technical proposals: the non-stationary fluctuating wind speed Forecasting Methodology that the present invention is based on EMD-ELM comprises the following steps:
The first step: become autoregressive moving-average model simulation when utilizing and generate non-stationary fluctuating wind speed sample, fluctuating wind speed sample is divided into training set, test set two parts, and adopts Matlab to samples normalization process;
Nonlinear for this non-stationary fluctuating wind signal decomposition is one group of stable state and linear sequence sets, i.e. intrinsic mode function by second step: carry out empirical mode decomposition process to the time series of this non-stationary fluctuating wind speed sample;
3rd step: phase space reconfiguration is carried out to this group intrinsic mode function component, and set up corresponding extreme learning machine forecast model respectively according to their respective features, study prediction is carried out to this non-stationary fluctuating wind speed time series;
4th step: predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed that just can obtain the non-stationary fluctuating wind of this point, simultaneously by the non-stationary fluctuating wind speed Comparative result of test sample book and prediction, the average error of computational prediction wind speed and actual wind speed, root-mean-square error and related coefficient, evaluate the validity of this method.
Preferably, in the above-mentioned first step, in time, becomes autoregressive moving-average model simulation m and ties up non-stationary fluctuating wind speed and be expressed as following formula:
U ( t ) = Σ i = 1 p A i ( t ) U ( t - i Δ t ) + Σ j = 0 q B j ( t ) X ( t - j Δ t )
Wherein, U (t) is zero-mean nonstationary random process vector, A i(t) for time become autoregressive coefficient matrix, B j(t) for time become slip regression coefficient matrix, p is Autoregressive, and q is slip regression order, and to be variance be X (t) 1, the white noise sequence of normal distribution.
Preferably, in the above-mentioned first step, time become the Autoregressive p=4 of autoregressive moving-average model, slip regression order q=1; Simulation points is positioned at along downburst moving direction and distance downburst thunderstorm center 3500m; Downburst Wind speed model adopts Oseguera and Bowles mean wind speed model, the vertical distributed model of Vicroy, maximum wind velocity V in vertical distribution wind speed max=80m/s, residing height and position Z max=67m; The radial maximum wind velocity V of certain At The Height in wind speed field r, max=47m/s, with downburst central horizontal distance r max=1000m, radical length scale-up factor R r=700m; Thunderstorm intensity time variations following formula represents:
Π = t / 5 , ( 0 ≤ t ≤ 5 min ) e - ( t - 5 ) / 20 , ( t > 5 min )
In formula, ∏ represents thunderstorm intensity, and t represents the time, and e is natural constant, downburst point-to-point speed V o=8m/s; Upper cut-off frequency is 2 π rad, N=2^11, consider that downburst self moves, simulated time interval of delta t=0.5s, simulation duration is 1000s, totally 2000 sample points simultaneously.
Preferably, in described second step, the time series of the non-stationary fluctuating wind speed sample obtained being carried out empirical mode decomposition process, is one group of stable state and linear sequence sets, i.e. intrinsic mode function by nonlinear for this non-stationary fluctuating wind signal decomposition.
Preferably, in described 3rd step, this group intrinsic mode function component obtained is carried out phase space reconfiguration, just can determine the basic structure of extreme learning machine according to the dimension of sample vector after phase space reconfiguration and the extreme learning machine hidden neuron number of specifying; Then set up corresponding extreme learning machine forecast model respectively according to this group intrinsic mode function component feature separately, study prediction is carried out to this non-stationary fluctuating wind speed time series; Finally predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed that just can obtain the non-stationary fluctuating wind of this point.
Preferably, empirical mode decomposition non-stationary Wind Velocity History U (t) be expressed as intrinsic mode function component and add final surplus r n(t), as shown in the formula:
U ( t ) = Σ j = 1 n c j ( t ) + r n ( t )
In formula, c jt () represents a jth intrinsic mode function component, n represents that empirical mode decomposition becomes the quantity of intrinsic mode function, r nt () represents surplus.
The non-stationary fluctuating wind speed Forecasting Methodology that the present invention is based on EMD-ELM has following advantage: non-linear and non-stationary for each spatial point fluctuating wind speed, and forecast model has good learning ability for the intrinsic mode function component after empirical mode decomposition and training error is little.Meanwhile, adopt extreme learning machine to predict, guarantee the accuracy that non-stationary fluctuating wind speed is predicted.Show according to operation result, predict that the non-stationary fluctuating wind speed that obtains and actual non-stationary fluctuating wind speed coincide very well based on EMD-ELM method, can as a kind of effective ways of single-point non-stationary fluctuating wind speed prediction.
Accompanying drawing explanation
Fig. 1 be based on time become 20 meters of At The Height fluctuating wind speed sample schematic diagram that autoregressive moving-average model method simulates;
Fig. 2 is the empirical mode decomposition schematic diagram of 20 meters of At The Height non-stationary fluctuating wind speed analog samples;
Fig. 3 is based on EMD-ELM non-stationary fluctuating wind speed Forecasting Methodology design framework figure schematic diagram;
Fig. 4 is based on extreme learning machine fluctuating wind speed Forecasting Methodology process flow diagram;
Fig. 5 is that 20 meters of At The Height EMD-ELM prediction of wind speed and actual wind speed contrast schematic diagram;
Fig. 6 is that 20 meters of At The Height EMD-ELM prediction of wind speed and actual wind speed autocorrelation function contrast schematic diagram;
Fig. 7 is that 20 meters of At The Height EMD-ELM prediction of wind speed and actual wind speed power spectral density function contrast schematic diagram.
Embodiment
Design of the present invention is as follows: first carry out empirical mode decomposition to the non-stationary fluctuating wind speed of this point, is decomposed into a series of component relatively stably; Then each component is predicted respectively, adopt extreme learning machine to become autoregressive moving-average model in time to the method that each component is predicted; Predicting the outcome to superpose and obtain final prediction of wind speed finally to each component of this point.
The non-stationary fluctuating wind speed Forecasting Methodology that the present invention is based on EMD-ELM comprises the steps:
The first step, becomes autoregressive moving-average model (TARMA) simulation and generates non-stationary fluctuating wind speed sample, fluctuating wind speed sample is divided into training set, test set two parts, and adopts Matlab to samples normalization process when utilizing;
In the described first step, in time, becomes autoregressive moving-average model simulation m and ties up non-stationary fluctuating wind speed and be expressed as following formula (1):
U ( t ) = Σ i = 1 p A i ( t ) U ( t - i Δ t ) + Σ j = 0 q B j ( t ) X ( t - j Δ t ) - - - ( 1 )
In formula, U (t) is zero-mean nonstationary random process vector, A i(t) for time become autoregressive coefficient matrix, B j(t) for time become slip regression coefficient matrix, p is Autoregressive, and q is slip regression order, and to be variance be X (t) 1, the white noise sequence of normal distribution.
Time become the Autoregressive p=4 of autoregressive moving-average model, slip regression order q=1.Simulation points is positioned at along downburst moving direction and distance downburst thunderstorm center 3500m.Downburst Wind speed model adopts Oseguera and Bowles mean wind speed model, the vertical distributed model of Vicroy, maximum wind velocity V in vertical distribution wind speed max=80m/s, residing height and position Z max=67m; The radial maximum wind velocity V of certain At The Height in wind speed field r, max=47m/s, with downburst central horizontal distance r max=1000m, radical length scale-up factor R r=700m; Thunderstorm intensity time variations following formula (2) represents:
Π = t / 5 , ( 0 ≤ t ≤ 5 min ) e - ( t - 5 ) / 20 , ( t > 5 min ) - - - ( 2 )
In formula, ∏ represents thunderstorm intensity, and t represents the time, and e is natural constant, downburst point-to-point speed V 0=8m/s.Upper cut-off frequency is 2 π rad, N=2 11, consider that downburst self moves, simulated time interval of delta t=0.5s, simulation duration is 1000s, totally 2000 sample points simultaneously.The time spectrum of the non-stationary fluctuating wind speed simulation of 20m At The Height as shown in Figure 1.
Second step: empirical mode decomposition (EMD) process is carried out to the time series of this non-stationary fluctuating wind speed sample, be one group of stable state and linear sequence sets by nonlinear for this non-stationary fluctuating wind signal decomposition, i.e. intrinsic mode function (IMF);
That empirical mode decomposition non-stationary Wind Velocity History U (t) is expressed as intrinsic mode function component (IMFs) and add final surplus r n(t), as shown in the formula (3):
U ( t ) = Σ j = 1 n c j ( t ) + r n ( t ) - - - ( 3 )
Wherein, c jt () represents a jth intrinsic mode function component, n represents that empirical mode decomposition becomes the quantity of intrinsic mode function, r nt () represents surplus.
Empirical mode decomposition is carried out to the non-stationary fluctuating wind speed at 20m place, obtains intrinsic mode function component and the surplus of this spatial point, as shown in Figure 2.
3rd step: phase space reconfiguration is carried out to this group intrinsic mode function component, and set up corresponding extreme learning machine forecast model respectively according to their respective features, study prediction is carried out to this non-stationary fluctuating wind speed time series;
This group intrinsic mode function component obtained is carried out phase space reconfiguration, just can determine the basic structure of extreme learning machine according to the dimension of sample vector after phase space reconfiguration and the extreme learning machine hidden neuron number of specifying; Then set up corresponding extreme learning machine forecast model respectively according to this group intrinsic mode function component feature separately, study prediction is carried out to this non-stationary fluctuating wind speed time series; Finally predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed that just can obtain the non-stationary fluctuating wind of this point.
Be specially, first, consider that research object of the present invention is point speed in units of 0.5 second, sample number is 2000 time points altogether 1000s, thus choose front 1000 time points altogether 500s form training set, and then rear 1000 time points are total to 500s as test set.Then, in order to reflect this non-stationary fluctuating wind speed sample Evolution in time, carry out phase space reconfiguration to the intrinsic mode function component that this non-stationary fluctuating wind speed obtains after empirical mode decomposition, time delay τ=1 chosen during reconstruct, Embedded dimensions is m=10.Because Embedded dimensions is 10, so the sample of training set is 990 10 dimensional vectors, test set is 1000 10 dimensional vectors.Therefore, the structure of extreme learning machine is: input layer 10 neurons, hidden layer 20 neurons, output layer are 1 neuron, so just determine the structure of extreme learning machine non-stationary fluctuating wind speed forecast model.Last just with this ELM forecast model, decompose through EMD the IMFs produced to this non-stationary fluctuating wind speed and carry out study prediction, the process flow diagram of the Forecasting Methodology of ELM as shown in Figure 4.
4th step: predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed that just can obtain the non-stationary fluctuating wind of this point, simultaneously by the non-stationary fluctuating wind speed Comparative result of test sample book and prediction, the average error (AE) of computational prediction wind speed and actual wind speed, root-mean-square error (RMSE) and related coefficient (R), evaluate the validity of this method.In addition, the result that EMD-ELM predicts can be contrasted with predicting the outcome of EMD-BPNN.
Matching performance for extreme learning machine is subject to the impact of network structure to a certain extent, this method is used for the extreme learning machine model of forecasting wind speed, input layer number is the input saturation dimension of phase space reconfiguration, output layer has a neuron, what unique needs were specified is hidden layer neuron number, and this method adopts the extreme learning machine structure comprising 20 hidden neurons.
Fig. 5 is 20 meters of At The Height prediction of wind speed and the comparing of actual wind speed; Fig. 6 is 20 meters of At The Height EMD-ELM prediction of wind speed and the comparing of actual wind speed autocorrelation function; Fig. 7 is 20 meters of At The Height EMD-ELM prediction of wind speed and the comparing of actual wind speed power spectral density function.The average error (AE) of computational prediction wind speed and actual wind speed, root-mean-square error (RMSE) and related coefficient (R), evaluate validity of the present invention.
Step is above that the calculation procedure based on empirical mode decomposition and extreme learning machine fluctuating wind speed Forecasting Methodology worked out based on Matlab platform carries out analysis & verification.The evaluation index that predicts the outcome of EMD-ELM and EMD-BPNN is shown in Table 1.
Table 1 predicts the outcome each evaluation index table
Evaluation index Forecasting Methodology EMD-BPNN EMD-ELM
AE 0.6209 0.5617
RSME 1.1940 0.9374
R 0.9972 0.9977
Above step can reference diagram 3, gives implementing procedure of the present invention intuitively.Analysis result shows, and EMD-ELM and the EMD-BPNN related coefficient that predicts the outcome all is greater than 0.99 and namely illustrates there is very strong correlation, and square error shows EMD-ELM and predicts the outcome and better converge on actual wind speed.And, can find out that predicting the outcome of EMD-ELM is all better than EMD-BPNN in average error, root-mean-square error and related coefficient these three, so the estimated performance of EMD-ELM to non-stationary fluctuating wind speed is more excellent, the travelling speed of extreme learning machine is far faster than reverse transmittance nerve network simultaneously.The present invention is that the prediction of non-stationary fluctuating wind speed provides that a speed is faster, the method for better effects if.
Above-described specific embodiment; the technical matters of solution of invention, technical scheme and beneficial effect are further described; be understood that; the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (6)

1., based on a non-stationary fluctuating wind speed Forecasting Methodology of EMD-ELM, it is characterized in that, it comprises the following steps:
The first step: become autoregressive moving-average model simulation when utilizing and generate non-stationary fluctuating wind speed sample, fluctuating wind speed sample is divided into training set, test set two parts, and adopts Matlab to samples normalization process;
Nonlinear for this non-stationary fluctuating wind signal decomposition is one group of stable state and linear sequence sets, i.e. intrinsic mode function by second step: carry out Empirical Mode Decomposition Algorithm process to the time series of this non-stationary fluctuating wind speed sample;
3rd step: phase space reconfiguration is carried out to this group intrinsic mode function component, and set up corresponding extreme learning machine forecast model respectively according to their respective feature, study prediction is carried out to this non-stationary fluctuating wind speed time series;
4th step: predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed of the non-stationary fluctuating wind just obtaining this point, simultaneously by the non-stationary fluctuating wind speed Comparative result of test sample book and prediction, the average error of computational prediction wind speed and actual wind speed, root-mean-square error and related coefficient, evaluate the validity of this method.
2. the non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM according to claim 1, is characterized in that, in the described first step, time becomes autoregressive moving-average model simulation m and ties up non-stationary fluctuating wind speed and be expressed as following formula:
U ( t ) = Σ i = 1 p A i ( t ) U ( t - i Δ t ) + Σ j = 0 q B j ( t ) X ( t - j Δ t )
In formula, U (t) is zero-mean nonstationary random process vector, A i(t) for time become autoregressive coefficient matrix, B j(t) for time become slip regression coefficient matrix, p is Autoregressive, and q is slip regression order, and to be variance be X (t) 1, the white noise sequence of normal distribution.
3. the non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM according to claim 1, is characterized in that, in the described first step, time become the Autoregressive p=4 of autoregressive moving-average model, slip regression order q=1; Simulation points is positioned at along downburst moving direction and distance downburst thunderstorm center 3500m; Downburst Wind speed model adopts Oseguera and Bowles mean wind speed model, the vertical distributed model of Vicroy, maximum wind velocity V in vertical distribution wind speed max=80m/s, residing height and position Z max=67m; The radial maximum wind velocity V of certain At The Height in wind speed field r, max=47m/s, with downburst central horizontal distance r max=1000m, radical length scale-up factor R r=700m; Thunderstorm intensity time variations following formula represents:
Π = t / 5 , ( 0 ≤ t ≤ 5 min ) e - ( t - 5 ) / 20 , ( t > 5 min )
In formula, ∏ represents thunderstorm intensity, and t represents the time, and e is natural constant, downburst point-to-point speed V o=8m/s; Upper cut-off frequency is 2 π rad, N=2^11, consider that downburst self moves, simulated time interval of delta t=0.5s, simulation duration is 1000s, totally 2000 sample points simultaneously.
4. the non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM according to claim 1, it is characterized in that, in described second step, the time series of the non-stationary fluctuating wind speed sample obtained is carried out empirical mode decomposition process, be one group of stable state and linear sequence sets, i.e. intrinsic mode function by nonlinear for this non-stationary fluctuating wind signal decomposition.
5. the non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM according to claim 1, is characterized in that, in described second step, that empirical mode decomposition non-stationary Wind Velocity History U (t) is expressed as intrinsic mode function component and add final surplus r n(t), as shown in the formula:
U ( t ) = Σ j = 1 n c j ( t ) + r n ( t )
Wherein, c jt () represents a jth intrinsic mode function component, n represents that empirical mode decomposition becomes the quantity of intrinsic mode function, r nt () represents surplus.
6. the non-stationary fluctuating wind speed Forecasting Methodology based on EMD-ELM according to claim 1, it is characterized in that, in described 3rd step, this group intrinsic mode function component obtained is carried out phase space reconfiguration, just can determine the basic structure of extreme learning machine according to the dimension of sample vector after phase space reconfiguration and the extreme learning machine hidden neuron number of specifying; Then set up corresponding extreme learning machine forecast model respectively according to this group intrinsic mode function component feature separately, study prediction is carried out to this non-stationary fluctuating wind speed time series; Finally predicting the outcome of this group intrinsic mode function component is carried out superposing the prediction of wind speed that just can obtain the non-stationary fluctuating wind of this point.
CN201510556369.5A 2015-09-02 2015-09-02 Non-stationary fluctuating wind speed forecasting method based on EMD-ELM Pending CN105205495A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510556369.5A CN105205495A (en) 2015-09-02 2015-09-02 Non-stationary fluctuating wind speed forecasting method based on EMD-ELM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510556369.5A CN105205495A (en) 2015-09-02 2015-09-02 Non-stationary fluctuating wind speed forecasting method based on EMD-ELM

Publications (1)

Publication Number Publication Date
CN105205495A true CN105205495A (en) 2015-12-30

Family

ID=54953167

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510556369.5A Pending CN105205495A (en) 2015-09-02 2015-09-02 Non-stationary fluctuating wind speed forecasting method based on EMD-ELM

Country Status (1)

Country Link
CN (1) CN105205495A (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105653817A (en) * 2016-01-20 2016-06-08 吴本刚 Substation with precise recognition function
CN106126906A (en) * 2016-06-22 2016-11-16 重庆科技学院 Short-term wind speed forecasting method based on C C Yu ELM
CN106529680A (en) * 2016-10-27 2017-03-22 天津工业大学 Multiscale extreme learning machine integrated modeling method based on empirical mode decomposition
CN107944613A (en) * 2017-11-17 2018-04-20 湖南大学 A kind of hazard forecasting method based on symbolic network
CN108710973A (en) * 2018-05-18 2018-10-26 武汉大学 Wind power forecasting method based on wind-powered electricity generation weather typing feature selecting
CN108846208A (en) * 2018-06-17 2018-11-20 华东交通大学 A kind of wind-load method for numerical simulation that wet downburst environment is rainy
CN109446697A (en) * 2018-11-08 2019-03-08 辽宁工程技术大学 A kind of mine wind speed fault branch diagnostic method based on ELM
CN110261305A (en) * 2019-06-18 2019-09-20 南京东南建筑机电抗震研究院有限公司 Based on across footpaths continuous bridge damnification recognition methods such as the multispan for influencing line
CN111368392A (en) * 2019-12-31 2020-07-03 重庆大学 Single-sample non-stationary wind speed simulation method based on MEMD and SRM
CN113625371A (en) * 2021-08-06 2021-11-09 国网安徽省电力有限公司 Grading detection method for high-frequency strong wind of non-stationary weather system
CN117498352A (en) * 2023-12-29 2024-02-02 西安热工研究院有限公司 Wind speed prediction method and device based on energy storage auxiliary black start capability

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030161501A1 (en) * 2002-02-23 2003-08-28 Imove Inc. Image distortion for gun sighting and other applications
CN102095538A (en) * 2011-02-25 2011-06-15 天津大学 Data demodulation method for polarization maintaining fiber stress sensing
CN103400210A (en) * 2013-08-13 2013-11-20 广西电网公司电力科学研究院 Short-term wind-speed combination forecasting method
CN104200102A (en) * 2014-09-03 2014-12-10 中国建筑设计咨询有限公司 CFD-based photovoltaic array wind load predicting method
US20150039538A1 (en) * 2012-06-01 2015-02-05 Mohamed Hefeeda Method for processing a large-scale data set, and associated apparatus
CN104573249A (en) * 2015-01-16 2015-04-29 上海大学 Time-variant ARMA model based non-stable wind speed simulation method
CN104573371A (en) * 2015-01-16 2015-04-29 上海大学 Optimal wavelet packet transformation and AR-SVM based non-stable wind speed prediction method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030161501A1 (en) * 2002-02-23 2003-08-28 Imove Inc. Image distortion for gun sighting and other applications
CN102095538A (en) * 2011-02-25 2011-06-15 天津大学 Data demodulation method for polarization maintaining fiber stress sensing
US20150039538A1 (en) * 2012-06-01 2015-02-05 Mohamed Hefeeda Method for processing a large-scale data set, and associated apparatus
CN103400210A (en) * 2013-08-13 2013-11-20 广西电网公司电力科学研究院 Short-term wind-speed combination forecasting method
CN104200102A (en) * 2014-09-03 2014-12-10 中国建筑设计咨询有限公司 CFD-based photovoltaic array wind load predicting method
CN104573249A (en) * 2015-01-16 2015-04-29 上海大学 Time-variant ARMA model based non-stable wind speed simulation method
CN104573371A (en) * 2015-01-16 2015-04-29 上海大学 Optimal wavelet packet transformation and AR-SVM based non-stable wind speed prediction method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张翌晖等: "基于集合经验模态分解和改进极限学习机的短期风速组合预测研究", 《电力系统保护与控制》 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105653817A (en) * 2016-01-20 2016-06-08 吴本刚 Substation with precise recognition function
CN106126906A (en) * 2016-06-22 2016-11-16 重庆科技学院 Short-term wind speed forecasting method based on C C Yu ELM
CN106529680B (en) * 2016-10-27 2019-01-29 天津工业大学 A kind of multiple dimensioned extreme learning machine integrated modelling approach based on empirical mode decomposition
CN106529680A (en) * 2016-10-27 2017-03-22 天津工业大学 Multiscale extreme learning machine integrated modeling method based on empirical mode decomposition
CN107944613A (en) * 2017-11-17 2018-04-20 湖南大学 A kind of hazard forecasting method based on symbolic network
CN108710973A (en) * 2018-05-18 2018-10-26 武汉大学 Wind power forecasting method based on wind-powered electricity generation weather typing feature selecting
CN108846208A (en) * 2018-06-17 2018-11-20 华东交通大学 A kind of wind-load method for numerical simulation that wet downburst environment is rainy
CN109446697A (en) * 2018-11-08 2019-03-08 辽宁工程技术大学 A kind of mine wind speed fault branch diagnostic method based on ELM
CN110261305A (en) * 2019-06-18 2019-09-20 南京东南建筑机电抗震研究院有限公司 Based on across footpaths continuous bridge damnification recognition methods such as the multispan for influencing line
CN111368392A (en) * 2019-12-31 2020-07-03 重庆大学 Single-sample non-stationary wind speed simulation method based on MEMD and SRM
CN111368392B (en) * 2019-12-31 2024-04-05 重庆大学 Single-sample non-stationary wind speed simulation method based on MEMD and SRM
CN113625371A (en) * 2021-08-06 2021-11-09 国网安徽省电力有限公司 Grading detection method for high-frequency strong wind of non-stationary weather system
CN113625371B (en) * 2021-08-06 2023-08-15 国网安徽省电力有限公司 Graded detection method for high-frequency strong wind of non-stationary weather system
CN117498352A (en) * 2023-12-29 2024-02-02 西安热工研究院有限公司 Wind speed prediction method and device based on energy storage auxiliary black start capability
CN117498352B (en) * 2023-12-29 2024-04-16 西安热工研究院有限公司 Wind speed prediction method and device based on energy storage auxiliary black start capability

Similar Documents

Publication Publication Date Title
CN105205495A (en) Non-stationary fluctuating wind speed forecasting method based on EMD-ELM
CN105138782A (en) Non-stationary pulse wind speed high-precision prediction method based on EEMD-ELM
Jiang Prediction of monthly mean daily diffuse solar radiation using artificial neural networks and comparison with other empirical models
Onat et al. Analysis of wind climate and wind energy potential of regions in Turkey
CN106650825A (en) Automotive exhaust emission data fusion system
CN105354363A (en) Fluctuation wind speed prediction method based on extreme learning machine
CN105243259A (en) Extreme learning machine based rapid prediction method for fluctuating wind speed
Higgins et al. Application of artificial intelligence to urban wind energy
CN104899432A (en) Kernel function combination-based PSO-LSSVM fluctuating wind speed prediction method
Huang et al. A new air quality forecasting model using data mining and artificial neural network
CN104951798A (en) Method for predicting non-stationary fluctuating wind speeds by aid of LSSVM (least square support vector machine) on basis of EMD (empirical mode decomposition)
CN104573249A (en) Time-variant ARMA model based non-stable wind speed simulation method
Ye et al. Probabilistic forecast of wind speed based on Bayesian emulator using monitoring data
CN104881563A (en) Chaotic characteristic analysis and non-linear prediction method of run-off
CN112100911A (en) Solar radiation prediction method based on deep BISLTM
Assas et al. Use of the artificial neural network and meteorological data for predicting daily global solar radiation in Djelfa, Algeria
CN115577436A (en) Combined deep learning method for solving wind-induced vibration response of uncertain structure
Chattopadhyay et al. Identification of the best hidden layer size for three-layered neural net in predicting monsoon rainfall in India
Stathopoulos Wind loads on low buildings: in the wake of Alan Davenport's contributions
Gupta et al. Short-term prediction of wind power density using convolutional LSTM network
Nguyen et al. Prediction of daily and monthly rainfall using a backpropagation neural Network
CN116976227B (en) Storm water increasing forecasting method and system based on LSTM machine learning
CN103870699B (en) Hydrodynamics flood routing analogy method based on double-deck asynchronous iteration strategy
Kisi et al. Wind speed prediction by using different wavelet conjunction models
Bashir et al. Decomposition and Harris hawks optimized multivariate wind speed forecasting utilizing sequence2sequence-based spatiotemporal attention

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20151230

RJ01 Rejection of invention patent application after publication