Tube-based robust model predictive control of multi-zone demand-controlled ventilation systems for energy saving and indoor air quality

Abstract

It is promising to apply model predictive control (MPC) scheme to demand-controlled ventilation (DCV) for the energy-efficient provision of indoor air quality (IAQ). However, the application of MPC in actual multi-zone DCV systems is challenged by the complexity of the ventilation duct network as well as the difficulty in handling uncertainties in actual systems. To tackle these issues, we proposed a novel robust model predictive control (MPC) strategy for multi-zone DCV systems. First, a data-driven model is established for ventilation duct network to represent relationships between airflow and damper angles. Then, with the above model and zone IAQ dynamics models, a two-layer tube-based MPC scheme is designed to efficiently handle uncertainties in actual systems: the first-layer MPC generates nominal state trajectories based on nominal systems without uncertainties while the second-layer MPC generates control actions to direct states of actual uncertain system to nominal trajectories. With the proposed strategy, the optimal trajectories of damper angles and fan pressure can be determined to minimize energy consumption while maintaining satisfying IAQ in the presence of uncertainties. Experimental results show that, in the presence of uncertainties, the proposed strategy can reduce IAQ cost by 10% and energy consumption by 14% compared with the strategy adopting conventional feedback control scheme. When compared with the baseline strategy without optimization, the proposed strategy can reduce IAQ cost by 38% and energy consumption by 30%. In addition, the proposed strategy can provide flexible trade-off between energy-efficiency and robustness against uncertainties based on designers’ interest.