Energy Consumption - Forecasting for Linköping and Helsingborg : A Comparison of SARIMA Modelling and Double Seasonal Holt-Winters Exponential Smoothing

• Main Author: Wöstmann, Corinna
• Format: Others
• Language: English
• Published: Linnéuniversitetet, Institutionen för matematik (MA) 2013
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-28318

Energy consumption forecasting is important for ecient planning of electricity production in order to minimize costs and also to ensure that the demand will be covered. In this thesis two methods are suggested: a combination of SARIMA modelling and regression, and forecasting by using the double seasonal Holt-Winters algorithm for exponential smoothing. Both methods are tested on energy consumption data for the regions Linkoping and Helsingborg provided by Bixia. Private and industry clients are regarded separately for regular days and holidays. The exponential smoothing method leads to good results for dierent forecasting horizons. Therefore an easier method, that is at least competitive to forecasts with SARIMA models, is found.