Deep Reinforcement Learning Based Energy-Efficient Heating Controller for Smart Buildings
Open Access
- Author:
- Gupta, Anchal
- Graduate Program:
- Engineering Science
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- November 07, 2019
- Committee Members:
- Youakim Badr, Thesis Advisor/Co-Advisor
Ashkan Negahban, Committee Member
Guanghua Qiu, Committee Member
Colin Neill, Program Head/Chair - Keywords:
- Deep reinforcement learning
Simulation
Occupant thermal comfort
Heating controller
HVAC - Abstract:
- Buildings account for 40% of the total energy consumption in the world out of which heating, ventilation, and air conditioning are the major contributors. Traditional heating controllers are inefficient due to lack of adaptability to dynamic conditions such as changing user preferences and extreme outside weather conditions. Therefore, it is necessary to design energy-efficient smart buildings which can improvise user comfort while reducing energy consumption. This research presents a Deep Reinforcement Learning (DRL) based heating controller to improve occupant comfort and minimize energy costs in smart buildings. At first, experiments are performed on one building with consideration of synthetic and real-world outside weather data. The DRL-based smart controller decides in realtime and outperforms the traditional heating controllers by improving user comfort between 15% - 30% and reducing energy costs between 5% - 12% in a simulated environment. Experiments are also performed on multiple zones, each having separate heating equipment. The performance is compared for when the multiple heaters are controlled centrally versus decentralized control, where each heater is controlled independently under various settings. As the number of zones and their complexity increase, decentralized controller performs better than the centralized controller. These results have significant practical implications for heating control in buildings consisting for many similar zones, such as office, classroom and apartment buildings as well as hotels and conference centers.