Title: Univariate forecasting of day-ahead hourly electricity demand in the northern grid of India

Authors: Sajal Ghosh

Addresses: Management Development Institutes, Room No. C – 10, P.O. Box No. 60, Mehrauli Road, Sukhrali, Gurgaon 122001, India

Abstract: Short-term electricity demand forecasts (minutes to several hours ahead) have become increasingly important since the rise of the competitive energy markets. The issue is particularly important for India as it has recently set up a power exchange (PX), which has been operating on day-ahead hourly basis. In this study, an attempt has been made to forecast day-ahead hourly demand of electricity in the northern grid of India using univariate time-series forecasting techniques namely multiplicative seasonal ARIMA and Holt-Winters multiplicative exponential smoothing (ES). In-sample forecasts reveal that ARIMA models, except in one case, outperform ES models in terms of lower RMSE, MAE and MAPE criteria. We may conclude that linear time-series models works well to explain day-ahead hourly demand forecasts in the northern grid of India. The findings of the study will immensely help the players in the upcoming power market in India.

Keywords: ARIMA; day-ahead hourly demand; electricity demand; exponential smoothing; India; univariate forecasting; power supply.

DOI: 10.1504/IJICBM.2009.027180

International Journal of Indian Culture and Business Management, 2009 Vol.2 No.6, pp.625 - 637

Published online: 17 Jul 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article