Keywords
K-means algorithm, data clustering, RNN neural network model, power load big data, forecasting method
Abstract
The existing methods ignore the clustering optimization process when predicting the short-term load of electricity, which leads to a long prediction time and a low accuracy of short-term load prediction. Therefore, a short-term load forecasting method based on improved K-means algorithm is proposed. This method uses the improved K-means algorithm to cluster the big data of power load, uses the training samples obtained after clustering to construct the RNN topology structure of the recurrent neural network. Then the optimal weights are set for the RNN neural network model to realize short-term forecast of the power load. The experimental results show that the proposed method has high forecasting efficiency and high short-term load forecasting accuracy.
DOI
10.19781/j.issn.1673-9140.2022.01.011
First Page
90
Last Page
95
Recommended Citation
XUN, Chao; CHEN, Bojian; WU, Xiangyu; XIANG, Kangli; LIN, Keyao; XIAO, Fen; and YI, Yang
(2022)
"Research on short-term power load forecasting method based on improved K-means algorithm,"
Journal of Electric Power Science and Technology: Vol. 37:
Iss.
1, Article 11.
DOI: 10.19781/j.issn.1673-9140.2022.01.011
Available at:
https://jepst.researchcommons.org/journal/vol37/iss1/11