An occupants’ diversity-aware discussion on the economic benefits of model predictive control in buildings

Abstract

The effectiveness of model predictive control (MPC) depends heavily on the flexible temperature band. In this study, we derived statistical distributions for the upper and lower bounds of acceptable temperature bands using survey data from the ASHRAE Global Thermal Comfort Database. A sampling technique was then applied to determine the acceptable temperature bands for thermal zones with varying occupant numbers. This approach accounts for occupants’ diversity in thermal preference and facilitates a consistent and unbiased assessment of MPC’s energy saving potential across different scenarios. Additionally, to enhance computational efficiency, we developed a linearized finite-difference building model with the same modeling capabilities and predictive accuracy as white-box models but compatible with linear programming for fast optimization. Using this model, we analyzed HVAC energy costs and organizational benefits of three MPC strategies—load shifting, peak shaving, and their combination—across 13 climate zones and three construction types. It was discovered that MPC reduces HVAC energy costs in all climates, with high absolute cost savings in hot climates and high cost-saving percentages in marine and cold climates. Across all scenarios, peak shaving is more effective than load shifting, and their combination achieves the highest cost savings.