Full Description

Natural ventilation is a promising passive technology to improve building energy performance and indoor air quality. However, the control of natural ventilation is a challenge in building technology and is often missing in building system advanced control design. This paper proposes the application of Model Predictive Control (MPC) and Reinforcement Learning (RL) control for winter natural ventilation control and evaluates both through on-site control experiments. Furthermore, this paper suggests the internal connection between MPC and RL control as simulation-based control, by transforming the MPC design into RL control design. This paper also highlights the RL control as a potential solver-free solution in deployment for building systems.