CN106374465A - GSA-LSSVM model-based short period wind electricity generation power prediction method - Google Patents
GSA-LSSVM model-based short period wind electricity generation power prediction method Download PDFInfo
- Publication number
- CN106374465A CN106374465A CN201610991088.7A CN201610991088A CN106374465A CN 106374465 A CN106374465 A CN 106374465A CN 201610991088 A CN201610991088 A CN 201610991088A CN 106374465 A CN106374465 A CN 106374465A
- Authority
- CN
- China
- Prior art keywords
- power
- wind speed
- data
- wind
- gsa
- 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
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000005611 electricity Effects 0.000 title abstract 4
- 238000010586 diagram Methods 0.000 claims description 12
- 230000005484 gravity Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 238000010845 search algorithm Methods 0.000 claims description 5
- 230000008030 elimination Effects 0.000 claims description 4
- 238000003379 elimination reaction Methods 0.000 claims description 4
- 238000005259 measurement Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000000630 rising effect Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000002474 experimental method Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 9
- 238000006243 chemical reaction Methods 0.000 abstract description 7
- 230000008859 change Effects 0.000 abstract description 5
- 230000009194 climbing Effects 0.000 abstract description 4
- 239000002245 particle Substances 0.000 description 22
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000001174 ascending effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 238000012886 linear function Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000000053 physical method Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005293 physical law Methods 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Wind Motors (AREA)
Abstract
The invention discloses a GSA-LSSVM model-based short period wind electricity generation power prediction method. Complex original wind speed-power data is reasonably classified into three types according to wind speed climbing gradients, a defect of an excessive calculation amount of a model can be made up for, characteristic features of the classified data can be reinforced, a conversion relation between wind speed and power can be well demonstrated, a phenomenon that the same speed corresponds to different power can be alleviated, an improved GSA can help alleviate a defect that a conventional GSA method can easily fall into a trap of a local minim value to a certain degree, and accuracy of short period power prediction of a wind power field can be improved; defects that a conventional prediction method is not high in generalization capability, the complex conversion relation between the wind speed and the power cannot be demonstrated via a single model, and abrupt winds speed change is not taken into consideration in a prediction model and other defects can be overcome; accuracy of real time prediction of actual electricity generation power of the wind power field can be improved, and the short period wind electricity generation power prediction method has good application prospects.
Description
Technical Field
The invention relates to the technical field of short-term wind power prediction, in particular to a short-term wind power prediction method based on a GSA-LSSVM model.
Background
With the rapid development of world economy and society, the traditional fossil energy faces the threat of exhaustion, renewable energy is highly valued, the instability of wind power generation is determined by the characteristics of wind fluctuation, randomness and the like along with the increase of the total amount of wind power installed equipment at home and abroad, and the problems of low wind power utilization rate and the like in the wind power industry are increasingly obvious, so that the accurate prediction of the real-time wind power is one of the important prerequisites for the stable and safe operation of a power grid.
At present, short-term wind power prediction methods can be divided into physical methods, statistical methods, artificial intelligence methods and combinations of the above methods. The physical method is based on a prediction result of Numerical Weather Prediction (NWP), considers physical factors such as terrain of a wind field and the like, and carries out power prediction according to a power curve of the wind turbine generator. The statistical method and the artificial intelligence method are based on a wind speed prediction time sequence provided by a wind power plant actual measurement time sequence or NWP (non-Newton P), and a prediction model is established through the time sequence data so as to predict the wind speed or the power generation power. Early statistical or intelligent methods used include single methods such as moving average autoregressive methods, artificial neural networks, Support Vector Machines (SVMs), and the like. Theoretically, the conversion from wind to power has a definite physical law, but in actual conditions, due to the influence of various objective factors, a single method is difficult to draw the conversion relationship from wind to power, and the single method has limitations, so wavelet decomposition, Genetic Algorithm (GA), particle swarm algorithm (PSO) and the like are introduced. Compared with the optimization algorithm, the Gravity Search Algorithm (GSA) is based on Newton's law of universal gravitation, and the acting force among particles does not need a propagation medium, so that the method has strong global property and convergence, which makes the method superior to the optimization algorithms such as GA and PSO.
Methods such as those described above all improve the prediction accuracy, but they are considered from the prediction method itself, and the influence of the wind speed rise and fall characteristics on the wind power prediction is less studied.
The concept of upwind and downwind is provided according to the lifting characteristic of the wind speed, a certain threshold value is set for marking a characteristic value for the wind speed at each moment, the training data dimension is increased, the rationality of the wind speed lifting characteristic in wind power prediction is verified by using three methods of an LSSVM (least squares support vector machine), an ELM (element-by-element model) and a GA-BP (genetic algorithm-BP), and compared with the method without considering the wind speed lifting characteristic, the prediction precision of the model can be obviously improved, and how to realize the description is the problem which needs to be solved at present.
Disclosure of Invention
The invention aims to solve the problems that in the prior art, research on the influence of wind speed lifting characteristics on wind power prediction is less, and prediction accuracy cannot be improved. The short-term wind power prediction method based on the GSA-LSSVM model overcomes the defects that the conventional prediction method cannot obtain better generalization capability, a single model cannot explain the complex conversion relation from wind speed to power, the sudden change of the wind speed is not considered to the prediction model and the like, improves the accuracy of the wind power plant in predicting the actual generated power in real time, and has good application prospect.
In order to achieve the purpose, the invention adopts the technical scheme that:
a short-term wind power prediction method based on a GSA-LSSVM model is characterized by comprising the following steps: comprises the following steps of (a) carrying out,
step (A), acquiring measured data (v (t) and p (t)) of a wind power plant, wherein t is sampling time, t is a natural number greater than 0, v (t) is measured wind speed, and p (t) is measured power;
step (B), performing data preprocessing on the actually measured data (v (t), p (t)) of the wind power plant, including abnormal data elimination and missing data filling, and obtaining denoised dataWherein,in order to remove the noise of the wind speed,is the denoised power;
step (C), denoising the data according to the formula (1)Classifying according to the time sequence to obtain new three groups of data, namely upwind data, downwind data and smooth data;
wherein RP is the gradient of wind speed rising or falling, Delta T is a time interval, and the values of the parameters RP and Delta T are determined by comparing the results of multiple experiments;
step (D), collecting the annual actual measurement wind speed power data of a plurality of fans into a wind speed power grade table, and setting the range of the wind speed as 0, vco]，vcoCutting out wind speed in m/s, dividing the wind speed into n grades, and obtaining the average value P of all power in the L grade according to the formula (2)LWherein L belongs to n, thereby obtaining a wind speed power curve which is the power of the average wind speed in the same level,
wherein p isiThe power corresponding to the ith wind speed on the L-th level is obtained, L represents the level of the wind speed power, and m represents the number of samples contained in the L-th level range;
step (E), drawing a wind speed and power scatter diagram of the classified upwind data, downwind data and steady data, comparing and observing the wind speed and power scatter diagram with a wind speed and power curve, determining the rule of upwind scatter points, downwind scatter points and the wind speed and power curve, determining that the wind speed rise and fall have influence on the wind power prediction precision, and obtaining that the upwind is slightly lower than the average level and the downwind is slightly higher than the average level;
selecting RBFs as kernel functions by applying an LSSVM model, determining punishment parameters gamma and RBF kernel hyper-parameters, optimizing the LSSVM model by adopting an improved gravity search algorithm GSA, determining optimal parameters, improving the prediction precision and convergence speed of the model, and forming an optimized GSA-LSSVM model;
and (G) taking the classified upwind data, downwind data and smooth data as the input of a GSA-LSSVM model, carrying out training by self-adapting to the parameters of the upwind, downwind and smooth three types, inputting the data to be predicted into the trained three types of models, recombining the obtained three types of power prediction results according to the element time sequence, and completing the prediction of the short-term wind power.
The short-term wind power prediction method based on the GSA-LSSVM model is characterized by comprising the following steps: step (D), the number of grades n ═ vcoD, d is step length of 0.05m/s, and wind speed range is [0,25 ]]The wind speed ranges of the respective levels are divided into [0, d ], [ d,2d ], [ (n-1) d, nd), n is 500, where nd is vco。
The invention has the beneficial effects that: according to the short-term wind power prediction method based on the GSA-LSSVM model, the complex original wind speed power data are reasonably classified into three types through the wind speed slope gradient, the defect that the calculated amount of the model is overlarge is made up, meanwhile, the classified data have stronger characteristics, the conversion relation from the wind speed to the power can be reflected better, the phenomenon that the same wind speed corresponds to different powers is improved, meanwhile, the defect that the traditional GSA method is prone to fall into the local minimum value is improved to a certain extent through the improved GSA, the accuracy of the short-term power prediction of the wind power plant is improved, and the method has a good application prospect.
Drawings
FIG. 1 is a flow chart of a short-term wind power prediction method based on a GSA-LSSVM model.
FIG. 2 is a schematic diagram illustrating comparison of data before and after denoising according to an embodiment of the present invention.
FIG. 3 is a schematic diagram of a wind speed power standard curve according to an embodiment of the present invention.
FIG. 4 is a schematic diagram illustrating a comparison between a scatter diagram and a wind speed power standard curve according to an embodiment of the present invention.
FIG. 5 is a schematic diagram comparing the average absolute error and the root mean square error of the GSA-LSSVM of the embodiment of the present invention with the conventional LSSVM and without classification of the ascending and descending wind.
Fig. 6 is a schematic diagram of the present invention in which the prediction curve gradually approximates the measured power curve.
Detailed Description
The invention will be further described with reference to the accompanying drawings.
The short-term wind power prediction method based on the GSA-LSSVM model makes up the defect of overlarge model calculation amount by reasonably classifying complicated original wind speed power data into three categories through wind speed climbing slopes, simultaneously has stronger data characteristics after classification, can better reflect the conversion relation of wind speed to power, improves the phenomenon of causing the same wind speed to correspond to different powers, simultaneously improves the defect of easy falling into local minimum value of the traditional GSA method to a certain extent by the improved GSA, thereby improving the accuracy of the short-term power prediction of the wind power plant, as shown in figure 1, comprises the steps of collecting wind field data, carrying out data denoising operation on the collected data, carrying out data decomposition according to the wind speed climbing slopes, further slowing down the complexity of the data and improving the characteristics of the data, and optimizing punishment parameters and kernel function hyper-parameters of the LSSVM model by using the improved GSA algorithm as a prediction model, the three types of decomposed data are respectively used as the input of the model, and finally the prediction results of the model are recombined, and the prediction process is completed,
step (A), acquiring measured data (v (t) and p (t)) of a wind power plant, wherein t is sampling time, t is a natural number greater than 0, v (t) is measured wind speed, and p (t) is measured power;
step (B), performing data preprocessing on the actually measured data (v (t), p (t)) of the wind power plant, including abnormal data elimination and missing data filling, and obtaining denoised dataWherein,in order to remove the noise of the wind speed,is the denoised power;
step (C), denoising the data according to the formula (1)Classifying according to the time sequence to obtain new three groups of data, namely upwind data, downwind data and smooth data;
wherein RP is the gradient of wind speed rising or falling, Delta T is a time interval, and the values of the parameters RP and Delta T are determined by comparing the results of multiple experiments;
step (D), collecting the annual actual measurement wind speed power data of a plurality of fans into a wind speed power grade table, and setting the range of the wind speed as 0, vco]，vcoCutting out wind speed in m/s, dividing the wind speed into n grades, and obtaining the average value P of all power in the L grade according to the formula (2)LWherein L belongs to n, thereby obtaining a wind speed power curve which is the power of the average wind speed in the same level,
wherein p isiThe power corresponding to the ith wind speed on the L-th level is obtained, L represents the level of the wind speed power, and m represents the number of samples contained in the L-th level range;
step (E), drawing a wind speed and power scatter diagram of the classified upwind data, downwind data and steady data, comparing and observing the wind speed and power scatter diagram with a wind speed and power curve, determining the rule of upwind scatter points, downwind scatter points and the wind speed and power curve, determining that the wind speed rise and fall have influence on the wind power prediction precision, and obtaining that the upwind is slightly lower than the average level and the downwind is slightly higher than the average level;
step (F), an LSSVM model is applied to select RBF as a kernel function, a penalty parameter gamma and an RBF kernel hyper-parameter are determined, and the change is adoptedThe gravity search algorithm GSA optimizes the LSSVM model, determines the optimal parameters, improves the prediction precision and the convergence speed of the model, and forms the optimized GSA-LSSVM model, which is specifically introduced as follows, RMSE of the prediction result of the LSSVM model is used as a fitness function fit, and the formula is as followsWherein N represents the number of samples, ytureRepresenting the actual power, ypredictRepresenting the predicted power;
the mass of the particles ismi(t) represents the mass of the particle i, mj(t) is a particlejMass of (1), fiti(t) represents the particle fitness value, best (t) best fitness value, worst (t) worst fitness value;
GSA particle bond forces are expressed asWhereinIs the magnitude of the attraction between particles i and j, G (t) is the attraction constant, Mi(t)、Mj(t) inertial masses of particles i, j, respectively, Rij(t) is the Euclidean distance between particles i and j, Rij(t)＝||xi(t),xj(t)||2Denotes a very small constant, prevents the denominator from being zero,Represents the ith particle in d dimension at time t,Denotes the jth particle, x, in d-dimension at time ti(t) t represents the ith particle at time t, xj(t) the jth particle at time t;
GSA particle accelerationThe degree is updated by the formulaWherein rand represents a random function,Representing the force between particles i and j in the d dimension at time t;
the gravity constant is the change process G (t) of the original exponential function G which is reduced along with the time0·e-α/TSubstituted by a linear function, the formula G (t) G0(1-t/T)，G0The traditional GSA algorithm G (T) adopts an exponential function form, the gravity coefficient attenuation speed is high, the global search capability of the GSA is weakened, the problem of premature convergence possibly occurs, a linear function is adopted to improve the gravity coefficient function in order to enable the later stage of the GSA to more accurately search an optimal solution area0Is particularly important, so that the dynamic selection mode is adopted to determine G0Is expressed ask is a coefficient of the number of the elements,andrespectively representing the value of the maximum position and the minimum position of the particle in the dimension space, so that the velocity and position updating formula of the particle i is as follows:calculating a fitness function value, and taking the optimal particles output by the improved GSA as a punishment parameter gamma and a kernel function hyperparameter sigma of the LSSVM model, so as to determine the optimal parameters, form the GSA-LSSVM model and improve the prediction precision and the convergence speed of the LSSVM model;
and (G) taking the classified upwind data, downwind data and smooth data as the input of a GSA-LSSVM model, carrying out training by self-adapting to the parameters of the upwind, downwind and smooth three types, inputting the data to be predicted into the trained three types of models, recombining the obtained three types of power prediction results according to the element time sequence, and completing the prediction of the short-term wind power.
According to the short-term wind power prediction method based on the GSA-LSSVM model, case analysis is carried out on wind speed power data collected every 5 minutes in the whole year 2015 of a certain wind power plant of Shanghai Chongming island for further explanation,
(1) acquiring measured data (v (t) and p (t)) of a wind power plant, wherein t is sampling time, t is a natural number greater than 0, v (t) is measured wind speed, and p (t) is measured power;
(2) preprocessing the data of (v (t), p (t)), including abnormal data elimination and missing data filling (denoising), and obtaining the denoised wind speed power
(3) According to the formulaTo denoised wind speed powerClassifying according to a time sequence, wherein RP is the gradient of wind speed rising or falling, delta T is a time interval, and determining a parameter V through comparison of a large number of experimental resultsrampAnd delta T, obtaining new three groups of data, namely upwind data, downwind data and smooth data;
(4) the LSSVM model is used for selecting the RBF as a kernel function, penalty parameters gamma and RBF kernel hyperparameters need to be determined, the time consumption of the traditional method adopting cross validation is long, the parameter selection effect is difficult to guarantee, the LSSVM model is optimized by adopting the improved Gravity Search Algorithm (GSA) introduced above, the optimal parameters are determined, and the prediction accuracy and the convergence rate of the model are improved;
(5) and respectively taking the classified three groups of new data as the input of a GSA-LSSVM model, adaptively adapting to the parameters of three types of upwind, downwind and peace, then training, inputting the data to be predicted into the trained three types of models, obtaining three types of power prediction results, and recombining according to the element time sequence.
In this embodiment, measured data (v (t), p (t)) of a wind farm is first preprocessed by a denoising method, as shown in fig. 2, the abscissa is wind speed, the ordinate is output power, and abnormal data are effectively processed by denoising, in this embodiment, wind speed power acquisition data of 30 fans in one year is selected, and a formula is usedThe measured data of the whole year is made into a wind speed and power registration table, as shown in table 2,
TABLE 2
Partial wind speed power rating table as shown in table 2
TABLE 2 partial wind speed power rating table
According to the comparison, a wind speed power standard curve is obtained, as shown in fig. 3, the wind speed power curve is the power of the average wind speed in the same grade, is relatively stable, and has a high wind speed power trend representative meaning.
Denoised dataCalculating formula according to wind speed climbing gradientThe decomposition is carried out into 3 types,and the power change condition under different wind speed characteristics is obviously reflected as shown in figure 4, and the upwind is slightly lower than the average level and the downwind is slightly higher than the average level;
RMSE of LSSVM prediction result as target function of GSABy the formulaExpressing the acting force between particles in GSA by formulaTo update the acceleration of the particles, G (t) G0·e-α/TUpdating the constant of gravity by formulaUpdating the speed and the position of the particles, finally calculating a fitness function value, and outputting an optimal value as a punishment parameter gamma and a kernel function hyperparameter sigma of the LSSVM model to form a GSA-LSSVM model;
taking the three types of separated data as training input data of the model, training a GSA-LSSVM model, separating the wind speed sequence corresponding to the power to be predicted into three types according to gradient, respectively inputting the three types of wind speed sequences into corresponding models, recombining prediction results of the three types of models according to the wind speed sequence, and recombining the prediction results according to a formulaCalculating root mean square error of predicted result, and method thereforAnd comparing the models to show the superiority of the model.
In order to analyze the beneficial effects of the invention, the prediction data of the method of the invention is respectively compared with the average absolute error and the root mean square error of the traditional LSSVM, and the GSA-LSSVM which does not classify ascending wind and descending wind, and the comparison result is shown in FIG. 5, the MAE and the RMSE of the method of the invention are obviously smaller than the LSSVM and the GSA-LSSVM, and the GSA-LSSVM is obviously superior to the LSSVM, which shows that the performance of the LSSVM is certainly improved by the GSA, and the classification prediction method further improves the model prediction accuracy; the prediction result of the RGSA-LSSVM model (the improved GSA optimizes the LSSVM model) is shown in FIG. 6, the prediction curve gradually approaches to the actually measured power curve, the obvious error has the tendency of gradual reduction, and the prediction result of the method is obvious through the comparative analysis of FIG. 6.
The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (2)
1. The short-term wind power prediction method based on the GSA-LSSVM model is characterized by comprising the following steps: comprises the following steps of (a) carrying out,
step (A), acquiring measured data (v (t) and p (t)) of a wind power plant, wherein t is sampling time, t is a natural number greater than 0, v (t) is measured wind speed, and p (t) is measured power;
step (B), performing data preprocessing on the actually measured data (v (t), p (t)) of the wind power plant, including abnormal data elimination and missing data filling, and obtaining denoised dataWherein,in order to remove the noise of the wind speed,is the denoised power;
step (C), denoising the data according to the formula (1)Classifying according to the time sequence to obtain new three groups of data, namely upwind data, downwind data and smooth data;
wherein RP is the gradient of wind speed rising or falling, Delta T is a time interval, and the values of the parameters RP and Delta T are determined by comparing the results of multiple experiments;
step (D), collecting the annual actual measurement wind speed power data of a plurality of fans into a wind speed power grade table, and setting the range of the wind speed as 0, vco]，vcoCutting out wind speed in m/s, dividing the wind speed into n grades, and obtaining the average value P of all power in the L grade according to the formula (2)LWhere L belongs to n, to obtain a wind speed power curveThe wind speed power curve is the power of the average wind speed in the same grade,
wherein p isiThe power corresponding to the ith wind speed on the L-th level is obtained, L represents the level of the wind speed power, and m represents the number of samples contained in the L-th level range;
step (E), drawing a wind speed and power scatter diagram of the classified upwind data, downwind data and steady data, comparing and observing the wind speed and power scatter diagram with a wind speed and power curve, determining the rule of upwind scatter points, downwind scatter points and the wind speed and power curve, determining that the wind speed rise and fall have influence on the wind power prediction precision, and obtaining that the upwind is slightly lower than the average level and the downwind is slightly higher than the average level;
selecting RBFs as kernel functions by applying an LSSVM model, determining punishment parameters gamma and RBF kernel hyper-parameters, optimizing the LSSVM model by adopting an improved gravity search algorithm GSA, determining optimal parameters, improving the prediction precision and convergence speed of the model, and forming an optimized GSA-LSSVM model;
and (G) taking the classified upwind data, downwind data and smooth data as the input of a GSA-LSSVM model, carrying out training by self-adapting to the parameters of the upwind, downwind and smooth three types, inputting the data to be predicted into the trained three types of models, recombining the obtained three types of power prediction results according to the element time sequence, and completing the prediction of the short-term wind power.
2. The short-term wind power prediction method based on the GSA-LSSVM model according to claim 1, characterized in that: step (D), the number of grades n ═ vcoD, d is step length of 0.05m/s, and wind speed range is [0,25 ]]The wind speed ranges of the respective levels are divided into [0, d ], [ d,2d ], [ (n-1) d, nd), n is 500, where nd is vco。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610991088.7A CN106374465B (en) | 2016-11-10 | 2016-11-10 | Short-term wind power forecast method based on GSA-LSSVM model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610991088.7A CN106374465B (en) | 2016-11-10 | 2016-11-10 | Short-term wind power forecast method based on GSA-LSSVM model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106374465A true CN106374465A (en) | 2017-02-01 |
| CN106374465B CN106374465B (en) | 2018-11-23 |
Family
ID=57894034
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610991088.7A Active CN106374465B (en) | 2016-11-10 | 2016-11-10 | Short-term wind power forecast method based on GSA-LSSVM model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106374465B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108053077A (en) * | 2017-12-28 | 2018-05-18 | 华中科技大学 | A kind of short-term wind speed forecasting method and system based on two type T-S fuzzy models of section |
| CN108154260A (en) * | 2017-12-15 | 2018-06-12 | 南京信息工程大学 | A kind of short-term wind power forecast method |
| CN108985490A (en) * | 2018-06-14 | 2018-12-11 | 内蒙古大学 | Wind-driven generator power curve modification method based on the fitting of sliding window characteristic |
| CN110880789A (en) * | 2019-12-06 | 2020-03-13 | 国网江苏省电力有限公司南通供电分公司 | Economic dispatching method for wind power and photovoltaic combined power generation system |
| CN111539577A (en) * | 2020-04-29 | 2020-08-14 | 南京信息工程大学 | Short-term wind power prediction method based on wind speed change rate and Gaussian process regression |
| CN112270439A (en) * | 2020-10-28 | 2021-01-26 | 国能日新科技股份有限公司 | Ultra-short-term wind power prediction method and device, electronic equipment and storage medium |
| CN116400028A (en) * | 2023-05-29 | 2023-07-07 | 湖南汇湘轩生物科技股份有限公司 | Essence quality detection method, system and medium based on smell sensor |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881705A (en) * | 2014-11-18 | 2015-09-02 | 中国石油大学(华东) | Method based on optimal combined model for predicting state trend of primary air fan |
-
2016
- 2016-11-10 CN CN201610991088.7A patent/CN106374465B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881705A (en) * | 2014-11-18 | 2015-09-02 | 中国石油大学(华东) | Method based on optimal combined model for predicting state trend of primary air fan |
Non-Patent Citations (3)
| Title |
|---|
| GUO PENGFEI , QI ZHIYUAN, HUANG WEI: "Short-Term Wind Power Prediction Based on Genetic Algorithm to Optimize RBF Neural Network", 《CONTROL AND DECISION CONFERENCE (CCDC)》 * |
| 张颖超，郭晓杰，叶小岭等: "一种短期风电功率集成预测方法", 《电力系统保护与控制》 * |
| 杜颖，卢继平，李青等: "基于最小二乘支持向量机的风电场短期风速预测", 《电网技术》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108154260A (en) * | 2017-12-15 | 2018-06-12 | 南京信息工程大学 | A kind of short-term wind power forecast method |
| CN108053077A (en) * | 2017-12-28 | 2018-05-18 | 华中科技大学 | A kind of short-term wind speed forecasting method and system based on two type T-S fuzzy models of section |
| CN108985490A (en) * | 2018-06-14 | 2018-12-11 | 内蒙古大学 | Wind-driven generator power curve modification method based on the fitting of sliding window characteristic |
| CN110880789A (en) * | 2019-12-06 | 2020-03-13 | 国网江苏省电力有限公司南通供电分公司 | Economic dispatching method for wind power and photovoltaic combined power generation system |
| CN110880789B (en) * | 2019-12-06 | 2022-04-22 | 国网江苏省电力有限公司南通供电分公司 | Economic dispatching method for wind power and photovoltaic combined power generation system |
| CN111539577A (en) * | 2020-04-29 | 2020-08-14 | 南京信息工程大学 | Short-term wind power prediction method based on wind speed change rate and Gaussian process regression |
| CN112270439A (en) * | 2020-10-28 | 2021-01-26 | 国能日新科技股份有限公司 | Ultra-short-term wind power prediction method and device, electronic equipment and storage medium |
| CN112270439B (en) * | 2020-10-28 | 2024-03-08 | 国能日新科技股份有限公司 | Ultra-short-term wind power prediction method and device, electronic equipment and storage medium |
| CN116400028A (en) * | 2023-05-29 | 2023-07-07 | 湖南汇湘轩生物科技股份有限公司 | Essence quality detection method, system and medium based on smell sensor |
| CN116400028B (en) * | 2023-05-29 | 2023-08-22 | 湖南汇湘轩生物科技股份有限公司 | Essence quality detection method, system and medium based on smell sensor |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106374465B (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106374465B (en) | Short-term wind power forecast method based on GSA-LSSVM model | |
| CN110006649B (en) | Bearing fault diagnosis method based on improved ant lion algorithm and support vector machine | |
| CN102880810B (en) | Wind power prediction method based on time sequence and neural network method | |
| CN106979126B (en) | Wind power generating set high wind speed section effective wind speed estimation method based on SVR | |
| CN106933778A (en) | A kind of wind power combination forecasting method based on climbing affair character identification | |
| CN103065202B (en) | Wind power plant ultrashort term wind speed prediction method based on combination kernel function | |
| CN107909211B (en) | Wind field equivalent modeling and optimization control method based on fuzzy c-means clustering algorithm | |
| CN111414717A (en) | XGboost-L ightGBM-based unit power prediction method | |
| CN105719029A (en) | Combined wind power prediction method based on wind speed fluctuation characteristic extraction | |
| CN108694482B (en) | Fractal theory and improved Least Square Support Vector Machine (LSSVM) tidal flow velocity prediction method | |
| CN102184337A (en) | Dynamic combination analysis method of new energy generating capacity influenced by meteorological information | |
| Wang et al. | A novel wind speed forecasting method based on ensemble empirical mode decomposition and GA-BP neural network | |
| CN104933483A (en) | Wind power forecasting method dividing based on weather process | |
| CN103278326A (en) | Method for diagnosing faults of wind generating set gear case | |
| CN105701572A (en) | Photovoltaic short-term output prediction method based on improved Gaussian process regression | |
| CN102509026A (en) | Comprehensive short-term output power forecasting model for wind farm based on maximum information entropy theory | |
| CN106529706A (en) | Support-vector-machine-regression-based method for predicting wind speed of wind power plant | |
| CN107045574A (en) | The low wind speed section effective wind speed method of estimation of wind power generating set based on SVR | |
| CN114648170A (en) | Reservoir water level prediction early warning method and system based on hybrid deep learning model | |
| CN110991701A (en) | Wind power plant fan wind speed prediction method and system based on data fusion | |
| CN115545070A (en) | Intelligent diagnosis method for unbalance-like bearing based on comprehensive balance network | |
| CN106786499B (en) | Based on the short-term wind power forecast method for improving AFSA optimization ELM | |
| CN115239029A (en) | Wind power prediction method and system considering power time sequence and meteorological dependent characteristics | |
| CN111488974A (en) | Deep learning neural network-based ocean wind energy downscaling method | |
| CN107527093B (en) | Wind turbine generator running state diagnosis method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230411 Address after: 4405-111, 4th Floor, Building 4, Shanxi Data Flow Ecological Park, No. 1 Huazhang North Street, Science and Technology Innovation City, Shanxi Transformation Comprehensive Reform Demonstration Zone, Taiyuan City, Shanxi Province, 030000 Patentee after: Shanxi Shendian Technology Co.,Ltd. Address before: 210044, No. 219, Ning six road, Pukou District, Jiangsu, Nanjing Patentee before: Nanjing University of Information Science and Technology |