CN110782071A - Method for predicting wind power by convolutional neural network based on time-space characteristic fusion - Google Patents
Method for predicting wind power by convolutional neural network based on time-space characteristic fusion Download PDFInfo
- Publication number
- CN110782071A CN110782071A CN201910914143.6A CN201910914143A CN110782071A CN 110782071 A CN110782071 A CN 110782071A CN 201910914143 A CN201910914143 A CN 201910914143A CN 110782071 A CN110782071 A CN 110782071A
- Authority
- CN
- China
- Prior art keywords
- time
- neural network
- convolutional neural
- space
- wind
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 41
- 230000004927 fusion Effects 0.000 title claims abstract description 9
- 238000002474 experimental method Methods 0.000 claims abstract description 8
- 230000008859 change Effects 0.000 claims abstract description 6
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000010801 machine learning Methods 0.000 abstract description 6
- 230000008034 disappearance Effects 0.000 abstract description 2
- 238000010248 power generation Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000002407 reforming Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Evolutionary Biology (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Public Health (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Wind Motors (AREA)
Abstract
The invention discloses a method for predicting wind power by a convolutional neural network based on time-space characteristic fusion, which comprises the following steps: step one, constructing a multi-feature model, extracting space-time features of wind-related information, and fusing the space-time features together to be used as the input of a convolutional neural network model; step two, constructing a convolutional neural network model, fitting the trend of the change of wind energy data, and performing wind energy prediction by combining the space-time characteristics obtained in the step one; respectively selecting two densely distributed and sparsely distributed areas of the turbine as data sets to carry out experiments; and step four, verifying the effectiveness of the invention by selecting the indexes of mean square error, root mean square error and peak signal-to-noise ratio. The invention is more effective and advanced in the aspect of extracting features. Meanwhile, compared with the convolutional neural network provided by other methods, the convolutional neural network constructed by the method is more advanced, can effectively solve the problem of gradient disappearance in machine learning, and is more suitable for the actual change condition of wind power.
Description
Technical Field
The invention relates to concepts and methods in the deep learning fields of convolutional neural networks, mean square error, root mean square error, peak signal-to-noise ratio and the like; in particular to a method for predicting wind power based on a convolutional neural network fused with space-time characteristics.
Background
Wind energy has been rapidly developed in power systems as an emerging important renewable energy source which can be developed and utilized on a large scale. Global Wind Energy Council (GWEC) predicts that by 2020, wind power generation capacity will reach 3.2 billion kilowatts. However, due to the influence of wind speed and wind direction, randomness and volatility of the turbine set are unavoidable, and therefore, accurate wind energy prediction is crucial to the safety and stability of the operation of the power system. The accurate wind energy prediction can enhance the controllability of wind power, ensure the stable operation of a power grid, reduce the cost of wind power generation and improve the capability of the power grid for receiving the wind power generation. Meanwhile, accurate wind energy prediction not only plays an important role in the reliability of the power system, but also provides a reference for decision making. In recent years, many scholars have worked to develop efficient and reliable wind speed and wind power prediction models, and have proposed many different approaches. The main methods currently used to predict wind power generation include: physical methods, statistical methods, and machine learning methods. The physical prediction model uses historical wind speed, terrain characterization data, and a host of meteorological data (atmospheric pressure, temperature, and humidity) to predict wind speeds for the site of interest. The statistical model is established on the basis of the statistical model, and the historical wind speed sequence is used for prediction. The machine learning method mainly adopts methods such as support vector machine regression (SVR), k-nearest neighbor regression (kNN), multi-layer perceptron neural network (MLP) long-short term memory network (LSTM) and the like to model the wind speed time sequence or the power time sequence so as to realize prediction. Machine learning methods effectively simplify the wind energy prediction problem, but in recent years there has been no significant improvement in the accuracy of the prediction.
Machine learning methods are often applied in short-term wind energy forecasting. Researchers make predictions of power or wind speed based on time series through regression models or neural networks. Common methods are SVR, kNN, multilayer perceptron network (MLP), long-short term memory neural network (LSTM), etc. With the rapid development of prediction techniques and the intensive research of wind speed prediction methods, in recent years, many new methods have been proposed: the signal decomposition algorithm is a popular idea for simplifying complex problems, and is mainly used for carrying out prediction pretreatment on an original wind speed sequence, and predicting precision, EMD (empirical mode decomposition), EEMD (integrated empirical mode decomposition) and other methods by eliminating the influence of outliers and noise in original data; the hybrid model combines a plurality of deep learning algorithms to improve the prediction capability of a single model, and due to the limitation of a single model, the hybrid model combining a plurality of single models can exert respective advantages at the same time. The methods essentially use time sequence data for modeling, and obtain higher precision through complex models, because the use of the complex models greatly increases the calculation cost of the complex models, and the models for extracting the characteristics of the complex models cannot reflect the space-time change of the wind power.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, and provides a method for predicting wind power by a convolutional neural network based on time-space characteristic fusion, which can effectively overcome the technical problem that the wind power can be predicted only by a single wind motor or only by one characteristic and the wind power cannot be accurately represented in the conventional wind power prediction method.
The purpose of the invention is realized by the following technical scheme:
the method for predicting the wind power based on the convolutional neural network with the time-space characteristic fusion comprises the following steps:
step one, constructing a multi-feature model, extracting space-time features of wind-related information, and fusing the space-time features together to be used as the input of a convolutional neural network model;
step two, constructing a convolutional neural network model, fitting the trend of the change of wind energy data, and performing wind energy prediction by combining the space-time characteristics obtained in the step one;
respectively selecting two densely distributed and sparsely distributed areas of the turbine as data sets to carry out experiments;
and step four, verifying the effectiveness of the invention by selecting the indexes of mean square error, root mean square error and peak signal-to-noise ratio.
Further, the specific steps in the first step are as follows: selecting a time interval of 10 minutes in a continuous period of time, constructing each type of wind-related information collected by the turbine into a two-dimensional image by a construction method of space-time characteristics, and sequencing the two-dimensional images according to a time sequence; in the combination of the two-dimensional image time sequences, each two-dimensional image independently represents space information, and the combination of more than one two-dimensional image time sequences represents time information; therefore, the combination of the two-dimensional image time series can comprehensively express the space-time information in the region; and finally, combining and connecting the two-dimensional image time sequences together to be used as the input of the constructed convolutional neural network model.
Further, the second step specifically comprises the following steps:
(301) after receiving the time sequence combination of the two-dimensional images, performing convolution and fully extracting the features;
(302) extracting depth features and reducing the size of an image through a pooling layer;
(303) performing multilayer convolution; after each layer of convolution, the output result at that time is stored;
(304) integrating the output results of each layer of convolution, connecting the output results stored after each layer of convolution by full connection, then connecting with a fully connected neural network, and mapping the deep features to the output end of each turbine;
(305) connected to a fully connected layer, reshaped into a two-dimensional image and map the output image with the pixels of the input image one by one.
Furthermore, in the third step, in the aspect of feature selection, a proposed convolutional neural network model is combined, and ten minutes is selected as an interval to extract historical features so as to perform an experiment.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
one method for improving the wind power prediction accuracy is to effectively extract the wind characteristics. The common method either only uses one kind of information or fails to effectively extract the spatiotemporal features of the wind-related information. The limitation of the methods to the extraction of the features limits the prediction accuracy of the methods. The multi-feature model for extracting features further integrates various information related to wind on the basis of extracting the space-time features of wind energy, discusses the features when the features are selected, and selects the optimal combination for prediction. Therefore, it is more efficient and advanced in extracting features. Meanwhile, compared with the convolutional neural network provided by other methods, the convolutional neural network constructed by the method is more advanced, can effectively solve the problem of gradient disappearance in machine learning, and is more suitable for the actual change condition of wind power.
Drawings
FIG. 1 is a schematic flow diagram of the present invention.
Fig. 2a, 2b, 2c, 2d, 2e, and 2f show probability density distributions of MSEs corresponding to different methods, respectively. Fig. 2a and 2b illustrate a method for predicting wind power of a single wind turbine, fig. 2a illustrates a prediction method using a kNN model, and fig. 2b illustrates a prediction method using an SVR model. FIG. 2c shows a prediction method using the E2E model, and FIG. 2d shows a prediction method using the FC-CNN model. FIG. 2e shows the model FB-CNN of the present invention. In FIG. 2f, the SVR model with the best prediction result for a single wind turbine is selected, and compared with the FC-CNN model with the best prediction result and the FB-CNN provided by the invention.
FIG. 3 is a graphical representation of the prediction error for each method on each turbine.
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a method for predicting wind power by a convolutional neural network based on time-space characteristic fusion, which fuses time-space characteristics of various wind-related information. As shown in fig. 1, an overall schematic diagram of a specific embodiment of the recommendation method of the present invention includes:
step S0101: a method of embedding the turbine into a grid of as small an area as possible is used, the algorithm first orders and discretizes the latitude and longitude coordinates to determine the shape of the scene, and then constructs the grid. Each turbine is ordered according to its horizontal and vertical coordinates, mapped to a corresponding grid cell, and the constructed grid is treated as a two-dimensional image.
Step S0102: because the constructed two-dimensional images represent the spatial distribution of wind power generation over time, concatenating multiple successive two-dimensional images may convey the process of spatial state over time. Each feature is individually constructed as a time series of two-dimensional images, which are finally connected together as input to a designed convolutional neural network model.
Step S0201: after receiving an input two-dimensional image sequence, performing convolution operation to retain the spatial information of an original input image, wherein the main task at this stage is to fully extract features, so that the number of channels in the feature image is rapidly increased, and then extracting depth features through a pooling layer to reduce the size of the image. A similar multi-layer convolution is then performed.
Step S0202: after each layer of convolution, the output result at that time is stored, the output of each layer of convolution layer is integrated, the convolution layers are connected by full connection, finally the convolution layers are connected with a full-connection neural network, and the deep layer characteristics are mapped to the output end of each turbine by fitting the complex function relation with the full-connection layer. And finally, connecting the output vector with a full connected layer, enabling the length of the output vector to be equal to the number of pixels in the input image, reforming the output vector into a two-dimensional image, and mapping the output and the pixels of the input image one by one.
Step S0301: the method firstly combines the power and other characteristics pairwise, uses the prediction time of 30 minutes, selects ten minutes as an interval to extract historical characteristics, and combines the proposed convolutional neural network to predict.
S0302, related wind power generation calculation formulas (4) and (5) are inquired, wherein A is a swept area, R is a radius, namely the length of a fan blade, V is a wind speed, Cp is a wind energy conversion rate value, the value is different according to different manufacturer technologies, D is an air density, the air density is reduced along with the rise of the altitude, η is a coefficient, four related characteristics are selected through the formulas to predict the wind energy, other characteristics are gradually added, and finally all the characteristics are used for prediction.
P＝1/2A*V3*Cp*D*η, (4)
A＝1/2π*R
2(5)
Step S0303: the experimental results of the present invention were compared to existing methods of prediction for a single turbine and two existing methods using convolutional neural networks (E2E model and FC-CNN model). And selecting a characteristic window (the number of extracted historical characteristics) to be 3,4 and 5 to perform a comparison experiment.
Step S0401: the accuracy of the prediction model was evaluated using Mean Square Error (MSE) and Root Mean Square Error (RMSE). The MSE calculation is shown in equation 1, RMSE is the square root of the mean square error, and the calculation is shown in equation (2), where X represents the sequence of true values, Y represents the sequence of predicted values, and n represents the length of the sequence.
The peak snr represents the ratio of the maximum possible power of a signal to the power of destructive noise affecting its accuracy of representation, and is often used as a measure of the quality of signal reconstruction in the field of image compression and the like, where MAXI is the maximum value representing the color of an image point, and if each sample point is represented by 8 bits, it is 255, as shown in equation (3).
Step S0401: the effectiveness of the invention is further illustrated by plotting tables, related graphs, such as fig. 2a to 3, the best results are obtained in both dense and sparse regions of the turbine.
The method constructs a multi-feature model, effectively fuses information related to wind, extracts corresponding space-time features, and takes the result as the input of the convolutional neural network. And the two data sets are used for verifying the prediction effect respectively, so that good experimental effects are obtained in dense and sparse areas. The accuracy of wind energy prediction is further improved, and the method is superior to the existing method. The power of the turbine set in a certain period can be well predicted, so that the pressure of peak load regulation and frequency modulation of a power system is relieved, and wind energy is fully utilized. Has huge value and wide prospect in practical application.
In the experiment, different combinations of power and other characteristics are used, and the convolutional neural network model provided by the invention is combined, ten minutes is selected as an interval to extract historical characteristics, and the experiment is carried out to select the combination with the best experimental result.
Experimental results show that the best prediction precision is achieved by using the combination of power, wind speed, air pressure, density and wind direction. Comparing the present invention with the experimental results of existing methods, fig. 2a, 2b, 2c, 2d, 2e, 2f show the probability density distribution of MSE corresponding to each method, respectively, with the horizontal axis representing the value of MSE and the vertical axis representing the Probability Density (PDF) function, and fig. 2a and 2b are methods of predicting a single turbine. It can be seen that the MSE distribution of the present invention is in a region of relatively small value compared to the method proposed by the present invention in fig. 2 e. The results are also better when compared to the two methods in fig. 2c and 2d, which use the E2E model and the FC-CNN model, respectively. Finally, in fig. 2f, the SVR and the convolutional neural network FC-CNN, which perform better single-point prediction, are put together with the method proposed herein for comparison, and it is obvious that the present invention has more areas in regions with smaller numerical values and better prediction effect. In order to more intuitively show the improvement of the effect of the invention, in fig. 3, 100 wind motors are selected, and the MSE of all the methods on the wind motors is drawn into a curve. In fact, compared with the SVR with the best single-point prediction result at present, the prediction precision is improved by 33.21% and 33.17%, and is improved by 36.55% and 35.75% compared with KNN. Compared with the conventional convolutional neural network method only using one characteristic, the prediction result of the method is more accurate, 1.736 and 1.916 are respectively achieved on two data sets, and the improvement is 9.38% and 7.13% compared with that of FC-CNN. The peak signal-to-noise ratio values for the proposed method on both data sets were also the highest as an image quality evaluation index, 45.730 and 45.307. The invention has further proved to be effective and the best results are obtained both in areas where the turbines are dense and in areas where the turbines are sparse.
The present invention is not limited to the above-described embodiments. The foregoing description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention, and the above specific embodiments are merely illustrative and not restrictive. Those skilled in the art can make many changes and modifications to the invention without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (4)
1. The method for predicting the wind power based on the convolutional neural network with the time-space characteristic fusion is characterized by comprising the following steps of:
step one, constructing a multi-feature model, extracting space-time features of wind-related information, and fusing the space-time features together to be used as the input of a convolutional neural network model;
step two, constructing a convolutional neural network model, fitting the trend of the change of wind energy data, and performing wind energy prediction by combining the space-time characteristics obtained in the step one;
respectively selecting two densely distributed and sparsely distributed areas of the turbine as data sets to carry out experiments;
and step four, verifying the effectiveness of the invention by selecting the indexes of mean square error, root mean square error and peak signal-to-noise ratio.
2. The method for predicting the wind power by the convolutional neural network based on the spatiotemporal characteristic fusion as claimed in claim 1, wherein the specific steps in the first step are as follows: selecting a time interval of 10 minutes in a continuous period of time, constructing each type of wind-related information collected by the turbine into a two-dimensional image by a construction method of space-time characteristics, and sequencing the two-dimensional images according to a time sequence; in the combination of the two-dimensional image time sequences, each two-dimensional image independently represents space information, and the combination of more than one two-dimensional image time sequences represents time information; therefore, the combination of the two-dimensional image time series can comprehensively express the space-time information in the region; and finally, combining and connecting the two-dimensional image time sequences together to be used as the input of the constructed convolutional neural network model.
3. The method for predicting the wind power based on the convolutional neural network with the fused spatiotemporal characteristics as claimed in claim 1, wherein the second step specifically comprises the following steps:
(301) after receiving the time sequence combination of the two-dimensional images, performing convolution and fully extracting the features;
(302) extracting depth features and reducing the size of an image through a pooling layer;
(303) performing multilayer convolution; after each layer of convolution, the output result at that time is stored;
(304) integrating the output results of each layer of convolution, connecting the output results stored after each layer of convolution by full connection, then connecting with a fully connected neural network, and mapping the deep features to the output end of each turbine;
(305) connected to a fully connected layer, reshaped into a two-dimensional image and map the output image with the pixels of the input image one by one.
4. The method for predicting the wind power by the convolutional neural network based on the spatio-temporal characteristic fusion as claimed in claim 1, wherein in the third step, in the aspect of feature selection, a proposed convolutional neural network model is combined, and ten minutes is selected as an interval to extract historical features so as to perform an experiment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910914143.6A CN110782071A (en) | 2019-09-25 | 2019-09-25 | Method for predicting wind power by convolutional neural network based on time-space characteristic fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910914143.6A CN110782071A (en) | 2019-09-25 | 2019-09-25 | Method for predicting wind power by convolutional neural network based on time-space characteristic fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110782071A true CN110782071A (en) | 2020-02-11 |
Family
ID=69384428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910914143.6A Pending CN110782071A (en) | 2019-09-25 | 2019-09-25 | Method for predicting wind power by convolutional neural network based on time-space characteristic fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110782071A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950704A (en) * | 2020-08-07 | 2020-11-17 | 哈尔滨工业大学 | Atmospheric temperature data generation method based on merging long-time and short-time memory networks |
| CN112949950A (en) * | 2021-04-29 | 2021-06-11 | 华北电力大学（保定） | Cluster wind power mapping prediction method based on multivariate space-time correlation matrix |
| CN113537573A (en) * | 2021-06-22 | 2021-10-22 | 中国农业大学 | Wind power operation trend prediction method based on dual space-time feature extraction |
| CN117495434A (en) * | 2023-12-25 | 2024-02-02 | 天津大学 | Electric energy demand prediction method, model training method, device and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108448610A (en) * | 2018-03-12 | 2018-08-24 | 华南理工大学 | A kind of short-term wind power prediction method based on deep learning |
| CN109117992A (en) * | 2018-07-27 | 2019-01-01 | 华北电力大学 | Ultra-short term wind power prediction method based on WD-LA-WRF model |
| CN109615146A (en) * | 2018-12-27 | 2019-04-12 | 东北大学 | A kind of wind power prediction method when ultrashort based on deep learning |
| CN109657839A (en) * | 2018-11-22 | 2019-04-19 | 天津大学 | A kind of wind power forecasting method based on depth convolutional neural networks |
-
2019
- 2019-09-25 CN CN201910914143.6A patent/CN110782071A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108448610A (en) * | 2018-03-12 | 2018-08-24 | 华南理工大学 | A kind of short-term wind power prediction method based on deep learning |
| CN109117992A (en) * | 2018-07-27 | 2019-01-01 | 华北电力大学 | Ultra-short term wind power prediction method based on WD-LA-WRF model |
| CN109657839A (en) * | 2018-11-22 | 2019-04-19 | 天津大学 | A kind of wind power forecasting method based on depth convolutional neural networks |
| CN109615146A (en) * | 2018-12-27 | 2019-04-12 | 东北大学 | A kind of wind power prediction method when ultrashort based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| SHUANG HU ET AL.: ""Short-Term Wind Power Prediction Based on Principal Compoent Analysis and Elman Artificial Neural Networks "" * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950704A (en) * | 2020-08-07 | 2020-11-17 | 哈尔滨工业大学 | Atmospheric temperature data generation method based on merging long-time and short-time memory networks |
| CN111950704B (en) * | 2020-08-07 | 2022-11-29 | 哈尔滨工业大学 | Atmospheric temperature data generation method based on merging long-time and short-time memory networks |
| CN112949950A (en) * | 2021-04-29 | 2021-06-11 | 华北电力大学（保定） | Cluster wind power mapping prediction method based on multivariate space-time correlation matrix |
| CN113537573A (en) * | 2021-06-22 | 2021-10-22 | 中国农业大学 | Wind power operation trend prediction method based on dual space-time feature extraction |
| CN113537573B (en) * | 2021-06-22 | 2023-11-28 | 中国农业大学 | Wind power operation trend prediction method based on double space-time feature extraction |
| CN117495434A (en) * | 2023-12-25 | 2024-02-02 | 天津大学 | Electric energy demand prediction method, model training method, device and electronic equipment |
| CN117495434B (en) * | 2023-12-25 | 2024-04-05 | 天津大学 | Electric energy demand prediction method, model training method, device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | A short-term load forecasting method based on GRU-CNN hybrid neural network model | |
| CN110782071A (en) | Method for predicting wind power by convolutional neural network based on time-space characteristic fusion | |
| CN109657839B (en) | Wind power prediction method based on deep convolutional neural network | |
| CN112529282A (en) | Wind power plant cluster short-term power prediction method based on space-time graph convolutional neural network | |
| CN108205717A (en) | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology | |
| CN110909919A (en) | Photovoltaic power prediction method of depth neural network model with attention mechanism fused | |
| CN106251001A (en) | A kind of based on the photovoltaic power Forecasting Methodology improving fuzzy clustering algorithm | |
| CN106295899B (en) | Wind power probability density Forecasting Methodology based on genetic algorithm Yu supporting vector quantile estimate | |
| CN110909911B (en) | Aggregation method of multidimensional time series data considering space-time correlation | |
| Moayyed et al. | A Cyber-Secure generalized supermodel for wind power forecasting based on deep federated learning and image processing | |
| CN116937579B (en) | Wind power interval prediction considering space-time correlation and interpretable method thereof | |
| CN111311001B (en) | Bi-LSTM network short-term load prediction method based on DBSCAN algorithm and feature selection | |
| CN114510869A (en) | Principal component analysis method and photovoltaic equipment power generation amount loss prediction method of Resnet network | |
| CN117233870B (en) | Short-term precipitation set forecasting and downscaling method based on multiple meteorological elements | |
| CN116524197B (en) | Point cloud segmentation method, device and equipment combining edge points and depth network | |
| CN116885691B (en) | Wind power climbing event indirect prediction method | |
| CN110826810B (en) | Regional rainfall prediction method combining spatial reasoning and machine learning | |
| CN115423810B (en) | Blade icing form analysis method for wind generating set | |
| CN116502074A (en) | Model fusion-based photovoltaic power generation power prediction method and system | |
| CN116167465A (en) | Solar irradiance prediction method based on multivariate time series ensemble learning | |
| CN115688982B (en) | Building photovoltaic data complement method based on WGAN and whale optimization algorithm | |
| Ajagunsegun et al. | Machine Learning-Based System for Managing Energy Efficiency of Public Buildings: An Approach towards Smart Cities | |
| Yang et al. | Investigating the predictability of photovoltaic power using approximate entropy | |
| CN114971007A (en) | Photovoltaic power generation prediction method and system based on multi-scale graph convolutional neural network | |
| CN115841167A (en) | Photovoltaic data prediction method based on multi-dimensional cross attention mechanism |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200211 |