A machine learning based heating and cooling load forecasting approach for DHC networks

Abstract

Space heating/cooling is one of the major contributors to the rapid growth in energy demand. In urban areas, district heating and cooling (DHC) systems are becoming increasingly popular for serving the thermal energy needs of consumers as it provides several advantages over the conventional electricity heating/cooling systems. But predicting heating and cooling loads in a DHC network is challenging because of their inherent differences with the conventional electrical networks. In this work, we have experimented with multiple machine learning and regression techniques on a real-life data set to model and forecast heating and cooling consumption load in a DHC network. We considered different types of weather parameters which are available easily. We demonstrate the performance of each algorithm with specific feature design for two different scenarios. Our experiments show that the error of forecasting can be reduced to strictly less than 5% using this methodology, which is better than most of the state-of-the-art results in this domain.