Short-term wind speed multi-step prediction method based on deep learning method

A deep learning and multi-step forecasting technology, applied in neural learning methods, biological neural network models, etc., can solve problems such as poor wind speed forecasting effect

Active Publication Date: 2013-11-27
哈尔滨工业大学人工智能研究院有限公司
View PDF1 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to solve the problem of poor wind speed prediction effect in the current wind speed prediction method. The present invention provides a short-term wind speed multi-step prediction method based on deep learning method

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Short-term wind speed multi-step prediction method based on deep learning method
  • Short-term wind speed multi-step prediction method based on deep learning method
  • Short-term wind speed multi-step prediction method based on deep learning method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0021] Specific implementation mode one: combine figure 1 Describe this embodiment, the short-term wind speed multi-step prediction method based on the deep learning method described in this embodiment,

[0022] Step 1: Based on the deep learning method, establish a deep neural network regression model with a multi-input multi-output structure;

[0023] Step 2: Using the layer-by-layer greedy method, combined with the recent actual wind speed data of the measured wind farm, the deep neural network regression model established in step 1 is trained, and through the nonlinear mapping function of the model, the sequence of the model is obtained through learning. The mapping relationship between them is used to determine the deep neural network regression model;

[0024] Step 3: According to the deep neural network regression model determined in step 2, perform multi-step prediction on the actual wind speed of the measured wind farm, and obtain the wind speed prediction result of ...

specific Embodiment approach 2

[0025] Embodiment 2: This embodiment is a further limitation of the short-term wind speed multi-step prediction method based on the deep learning method described in Embodiment 1.

[0026] In the described step one, based on the deep learning method, the method for setting up a deep neural network regression model with a multi-input multi-output structure is:

[0027] Constructing a deep neural network regression model with a multiple-input multiple-output structure by stacking restricted Boltzmann machines (RBMs),

[0028] The deep neural network regression model with the multi-input multi-output structure is an l-layer neural network, and the vector x=h 0 represents the original input, (h 1 ,...,h l-1 ) represents the input of the corresponding hidden layer, h l Represents the input of the output layer; its 1:l-1 hidden layer uses a sigmoid function and is composed of a restricted Boltzmann machine, and the top activation function uses a pure linear function;

[0029] Fo...

specific Embodiment approach 3

[0075] Embodiment 3: This embodiment is a further limitation of the short-term wind speed multi-step prediction method based on the deep learning method described in Embodiment 1.

[0076] Using the layer-by-layer greedy method, combined with the actual wind speed data of the measured wind farm, the deep neural network regression model established in step 1 is trained, and through the nonlinear mapping function of the model, the mapping relationship between the sequences of the model is obtained through learning. , to determine the process of deep neural network regression model as;

[0077] Step 21: Layer the deep neural network regression model established in step 1, from bottom to top, and then use the actual wind speed data of the measured wind farm to perform unsupervised training on the layer that inputs x;

[0078] Step 22: After the unsupervised training is over, supervised learning is used to fine-tune the deep neural network regression model:

[0079] In addition to...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a short-term wind speed multi-step prediction method based on a deep learning method, and relates to a short-term wind speed multi-step prediction method, which aims at solving the problem of poor wind speed prediction effect in an existing wind speed prediction method. The method comprises the following steps of 1, on the basis of the deep learning method, building a deep neural network regression model with a multiple input multiple output structure; 2. using of a gradual-layer greedy method, combining the recent actual wind speed data of a measured wind farm to train the built deep neural network regression model, and learning and obtaining the mapping relationship among the sequences of the obtained model through the non-linear mapping function of the model to determine the deep neural network regression model; and 3. according to the determined deep neural network regression model, carrying out multi-step prediction on the actual wind speed of the measured wind farm, so as to obtain the wind speed prediction result of the measured wind farm. The method is used for predicting the short-term wind speed of the wind farm.

Description

technical field [0001] The invention relates to a multi-step prediction method of short-term wind speed, in particular to a multi-step prediction method of short-term wind speed based on a deep learning method. Background technique [0002] Wind power generation is currently a hot topic in the research of the new energy power industry. Wind power itself has strong randomness and volatility. In recent years, a large number of large-scale wind farms have been put into operation in my country, which has brought many problems to the safe and stable operation of the power grid. The current installed capacity of wind power has begun to face the dilemma that the power grid cannot accommodate . Accurate wind farm power prediction is an important basis for solving this problem, which can help the power grid to formulate a reasonable dispatch plan, determine the spinning reserve, and ensure the operation of the power grid safely and economically. Wind is the power source of wind turb...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/08
Inventor 于达仁万杰胡清华刘金福郭钰锋苏鹏宇
Owner 哈尔滨工业大学人工智能研究院有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap