Short-term wind power prediction method based on multi-channel convolutional neural network and time convolutional network

Abstract

The invention discloses a short-term wind power prediction method based on a multi-channel convolutional neural network and a time convolutional network. The method comprises the steps of 1, extracting related features from wind power historical data and forming a sample set; 2, correcting the short-term prediction wind speed data in the sample set by using a multi-layer LSTM neural network model;3, performing normalization processing on the corrected sample set data, and dividing the sample set data into a training set and a test set; 4, respectively carrying out initialization setting on the three-channel CNNLSTM neural network and the TCN time convolution network, and respectively training the two networks by utilizing the training set; 5, respectively inputting the test set into the two neural network models, performing model fusion by using a weighted average method, and outputting a final prediction result; and 6, evaluating the prediction result according to various predefinedevaluation indexes. Experiments prove that the method has high accuracy and robustness for short-term wind power prediction.

Description

technical field [0001] The invention relates to the field of wind power, in particular to a short-term wind power prediction method based on a multi-channel convolutional neural network and a temporal convolutional network. Background technique [0002] Due to the non-renewability of fossil fuels and their greenhouse effect, countries all over the world have turned their attention to new renewable energy sources. Among them, due to its wide distribution and low cost, wind energy has been widely valued by countries all over the world. However, wind energy is inherently uncertain. With the gradual increase of wind power generation, the harm of wind power to the grid has become increasingly prominent. This is because the wind power is proportional to the cube of the wind speed, and small changes in wind speed will lead to large fluctuations in wind power, which will seriously affect the safety of the power grid and power quality, and will also increase the workload of power di...

AI Technical Summary

Problems solved by technology

This type of algorithm requires high basic knowledge of electricity and meteorology, and requires complex parameter adjustment. The calculation is relatively accurate, but the method is cumbersome and time-consuming, which is not conducive to large-scale promotion.

Images