“Model-Free Reinforcement Learning Based Control for Radiant Floor Heating Systems”, authored by Xu Han and Ali Malkawi, received the Best Paper Award at COBEE 2022 on July 27, 2022, in Montreal, Canada.
Only three papers out of 400 were selected for the honor at the meeting.
The award-winning research paper investigates the feasibility and strategies of using model-free, reinforcement learning-based control (RLC) for the slow-response radiant floor heating systems with a setback setting.
HouseZero third-floor radiant heat installation: Photo-CGBC
The study was conducted through a high-fidelity virtual testbed, which was developed and validated based on the third-floor lab in HouseZero. The effectiveness of the proposed model-free control strategy was demonstrated and benchmarked by comparing it with conventional rule-based control and model-based predictive control. Researchers concluded that the proposed model-free control strategy achieves a similar level of energy saving (19%) against the rule-based control as the model predictive control does. This study supports the feasibility and scalability of using data-driven approach for achieving ultra-energy efficiency.
