A Hybrid Method Based on Wavelet Analysis for Short-term Load Forecasting

Autor:	Ling Ji, Chenhao Niu
Rok vydání:	2012
Předmět:	Nonlinear system Wavelet Markov chain Computer Networks and Communications Hardware and Architecture Power load Computer science Load forecasting Power grid Volatility (finance) Markov model Algorithm
Zdroj:	Journal of Convergence Information Technology. 7:540-547
ISSN:	2233-9299 1975-9320
DOI:	10.4156/jcit.vol7.issue17.63
Popis:	Power load is a typical time series with the characters of nonlinearity and volatility. Accuracy load forecasting can help ensure the security of power grid operation. In this paper, a new hybrid method based on wavelet analysis for short-time load forecasting application is proposed. The key idea is that the original time series can be decomposed into various components with different frequencies though wavelet analysis. Then the low frequency part and high frequency components are predicted by RBF and Markov respectively. The final forecasting result is obtained by wavelet reconstruction. Though the experimental study, the hybrid method was found to perform better with higher accuracy compared to RBF and Markov models. The experimental results show the potential of this combined models based on wavelet analysis in the application of load forecasting.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::96068ac4d3bcc1313da4bf04347f8e3a https://doi.org/10.4156/jcit.vol7.issue17.63 Zobrazit plný text záznamu