human-centered control, personalized thermal preferences, model-predictive control, optimization, energy use
Model-based and model predictive control (MPC) based on personalized thermal preferences has the potential to minimize energy consumption and guarantee occupant thermal comfort. This paper presents the implementation of personalized controls that satisfy personalized thermal preferences in private offices. Two control strategies, a simple PID control and a real-time model predictive control, were implemented through the Building Management System in two adjacent identical office spaces. In each office, monitoring and control of indoor thermal conditions was achieved either by using a conventional wall thermostat or by using a low-cost wireless local sensing network -both integrated with the building management system. Thermal comfort and energy consumption performances were compared between MPC vs PID control, as well as local sensing vs conventional wall thermostat sensing. The results demonstrate the advantages of MPC over simple feedback control, in parallel with the advantages of local sensing over wall thermostats, in terms of energy efficiency and personalized thermal comfort.