Distributed air conditioning control in commercial buildings based on a physical-statistical approach

This paper presents a method to optimally control a commercial building air conditioning operation based on building thermal physics and human behavior. Control plans are developed for zone cooling and according to occupancy patterns at zone levels, and using a novel response model developed in this paper. The thermal response model is a statistical model, which is built on the basis of insights gained from physics-based model of zone thermal behavior. The control model attempts to optimize cooling system scheduling on the basis of occupancy patterns, price signal from utilities and human comfort and productivity. The optimization follows an on-demand routine, such that zone level air conditioning starts at a time epoch that is optimal according to zone thermal response and cooling requirements. We toss the term “pre-occupy” control. We also develop, what we call, “post-occupy” control, where the system shutdown follows thermal inertia and cooling requirements at the time that the zone is expected to become unoccupied. The thermal response modeling of a zone and the two control modes are considered major contributions of this article. EnergyPlus reference models are used to train our models and illustrative examples are presented.