Zone-level control algorithms based on occupancy information for energy efficient buildings

We examine the problem of how to use occupancy information of various fidelity to reduce the energy consumed in maintaining desired levels of thermal comfort and indoor air quality (IAQ) in commercial buildings. We focus on the zone-level control, where the control inputs to be decided are the supply air (SA) flow rate and the amount of reheat. We propose three control algorithms with varying information requirements: (i) POBOC, that requires long-horizon accurate prediction of occupancy and a model of the hygrothermal dynamics of the zone, (ii) OMBOC, that requires only occupancy measurement and a dynamic model, and (iii) Z-DCV, that requires only occupancy measurement. The first two strategies use a model predictive control framework to compute the optimal control inputs, while the third one is a pure feedback-based control strategy. Simulations with a calibrated model show that significant energy savings over a baseline controller, the kind usually used in existing buildings, is possible with the last two strategies, that is, even without occupancy prediction. Trade-offs between complexity and performance of the control algorithms are discussed.