Forecasting Heating Degree Days for Energy Demand Modeling

Abstract

Heating degree day (HDD) is a technical index taking into consideration outdoor temperature and average room temperature to describe the need for the heating energy requirements of buildings. HDD can be used to normalize the energy consumption of buildings with respect to heating since the amount of energy needed to heat a building in a given frequency is directly related to the number of heating degree days in that particular frequency. In order to understand the heating demand of the buildings, it is important to investigate the HDD patterns and to construct forecasting models. This study aims at constructing short-term forecast models by analysing the patterns of the HDD. Within this context, time series analysis was conducted by the monthly HDD data in France between 1974 and 2017. The performance of the models were assessed by the adjusted R value, residual sum of squares, the Akaike Information Criteria (AIC) and the Schwarz Information Criteria (SIC) as well as the analysis of the residuals. As a result, the most suitable model was determined as SARIMA (2,0,1)(1,0,1)12. The results of the study show that there is a potential to integrate time series models of HDD for short term load forecasting. © 2019 The Authors. Published by Budapest University of Technology and Economics & Diamond Congress Ltd. Peer-review under responsibility of the scientific committee of the Creative Construction Conference 2019.