Unlocking hidden energy efficiency potential in buildings using artificial intelligence algorithms for HVAC systems

This is the authors' abstract. We don't add key points for this paper.

In this paper, a supervisory control level that manages a HVAC system is described. The control is based on optimization strategies such as the modulation of the supply temperature subject also to indoor temperature, comfort boundaries, load balancing in Air Handling Units (AHU), and predictive control algorithms based on data such as forecasted occupancy. The proposed control system has been deployed in an airport, where it operates the AHUs by including external temperatures and flight timetable, and in a retrofitted Office Building where it operates heat pumps and boilers depending on room occupancy, indoor temperature and thermal inertia of each room. In the airport case study results have demonstrated an increase of internal comfort for terminal passengers regarding temperature and air ventilation given by a better HVAC management and, at the same time, a lower energy consumption ranging from 9% to 73%. In the retrofitted office building case study, a 28% reduction in electrical energy consumption and a 22% reduction in gas consumption were achieved allowing for an average supply water temperature reduction during wintertime up to 5.7 °C for heat pumps and 18 °C for the boilers and an average increase during summer up to 2.7 °C for chillers.