Hybrid modeling-based temperature and humidity adaptive control for a multi-zone HVAC system

In this paper, we study multi-zone air environment comfort and building energy conversation for the heating, ventilation, and air conditioning (HVAC) system. Each zone is equipped with the air-handling unit (AHU) and fan coil unit to regulate the supply air state for comfortable zone climate. It is known that the multi-zone air environment appears spatial and temporal variation due to the hygro-thermal interaction which is an extremely complex nonlinear dynamic process. In this study, hybrid models incorporating the first-principles model with full form dynamic linearization (FFDL)-based data-driven model have been proposed to precisely describe the multi-zone climate dynamics including unknown nonlinearity and uncertainty. Then model-free adaptive control (MFAC) scheme is designed for multi-zone climate control, which contains controller design, parameter estimation and reset mechanism to ensure system control performance. Finally, different case studies are conducted to test the performance of model prediction and system control for the multi-zone climate environment. Model validation shows that the zone climate prediction of the proposed hybrid model shows better agreement with actual measured data. The step response results display that the hybrid model-based MFAC can realize multi-zone climate control performance improvement in faster convergence without stable bias. By comparing with prior work, the hybrid model-based MFAC has the potential to get closer to actual situation in cost-effective multi-zone climate control.