Model-based predictive control for building energy management: Part II - Experimental validations

Indoor climate control of thermal comfort for humans in a residential or commercial building is a major component of building energy management. The goal of optimal temperature and humidity control is to ensure indoor comfort with minimal energy consumption. Model-Based Predictive Control (MBPC) is considered to be one of the most suited solutions to achieve this goal due to its ability to use building dynamics, occupancy schedule, and weather conditions for optimal control. The development and verification of MBPC have been discussed in the Part I  [1]. Here, to validate that MBPC achieves reduced energy consumption, while simultaneously satisfying comfort conditions, experiments are performed on a quarter scale shelter structure in a climate-controlled environmental chamber. The MBPC method is compared to three other control methods: conventional constant temperature setpoint control, scheduled control using a Honeywell smart thermostat, and scheduled control using Labview. Temperature variations and energy consumptions resulting from the four methods are analyzed. Compared to the three other methods, MBPC yields superior control performance with lowest energy consumption while still maintaining indoor thermal comfort. We also demonstrate that use of MBPC can reduce the number of sensors required for effective local control.