Title :
Electricity load forecasting based on weather variables and seasonalities: A neural network approach
Author :
Elias, R.S. ; Fang, Liping ; Wahab, M.I.M.
Author_Institution :
Dept. of Mech. & Ind. Eng., Ryerson Univ., Toronto, ON, Canada
Abstract :
Three models are presented to forecast electricity load based on weather variables, represented by heating degree days (HDD) and cooling degree days (CDD), and weekly and monthly seasonalities, expressed by the day of the week, a holiday, and the month of the year. The first model is a classical linear regression model that serves as a benchmark for this study. The second model is a feedforward neural network model, and the third is a nonlinear autoregressive time-lagged (NARx) neural network model for day-ahead electricity load forecasting. The results based on the mean absolute percentage error (MAPE) show that the third model outperforms the other two in forecasting day-ahead electricity load.
Keywords :
feedforward neural nets; load forecasting; power engineering computing; regression analysis; cooling degree days; day ahead electricity load forecasting; feedforward neural network; heating degree days; linear regression model; mean absolute percentage error; nonlinear autoregressive time lagged neural network; weather variable; Artificial neural networks; Electricity; Forecasting; Load modeling; Neurons; Predictive models; Training;
Conference_Titel :
Service Systems and Service Management (ICSSSM), 2011 8th International Conference on
Conference_Location :
Tianjin
Print_ISBN :
978-1-61284-310-0
DOI :
10.1109/ICSSSM.2011.5959472