The project aims to develop predictive controls to optimize fan coil control and window shading control using model-based and model-free computational strategies with the objective of improving energy efficiency and occupant comfort, including thermal comfort and visual comfort.
The goals of this project are to design and evaluate building controls for the fan coil unit and window shading system in the context of residential buildings.
We use simulations and laboratory experiments to design and validate the optimal building control designs. As for the solution strategies, we use linear programming, Q-learning, as well as policy gradient for finding and optimizing control actions of the fan coil unit and window shading system.
It is expected that the joint control is able to reduce the cooling and heating energy consumption by 10% to 30%, while maintaining an indoor temperature within a comfortable level.