Heat consumption forecasting using partial least squares, artificial neural network and support vector regression techniques in district heating systems

Effective management of district heating networks depends upon the correct forecasting of heat consumption during a certain period. In this work short-term forecasting for the amount of heat consumption is performed first to validate the three forecasting methods: partial least squares (PLS) method, artificial neural network (ANN), and support vector regression (SVR) method. Based on the results of short-term forecasting, one-week ahead forecasting was performed for the Suseo district heating network. Data of heat consumption and ambient temperature during January and February in 2007 and 2008 were employed as training elements. The heat consumption estimated was compared with actual one in the Suseo area to validate the forecasting models.

Authors and Affiliations

1. Department of Chemical Engineering, Hanyang University, Seoul, 133-791, Korea

   Tae Chang Park, Ui Sik Kim & Yeong Koo Yeo

2. Department of Chemical Engineering, Seoul National University of Technology, Seoul, 135-743, Korea

   Lae-Hyun Kim

3. Department of Civil Engineering, Hanyang University, Seoul, 133-791, Korea

   Byung Wan Jo

Corresponding author

Correspondence to Yeong Koo Yeo.

Reprints and permissions