Global optimization of VAV air conditioning system

Energy conversation is a goal of optimal control of Heating, ventilating and air-conditioning (HVAC) system. HVAC is a multi-variable, strongly coupled, nonlinear, time variant large-scale system composed of several subsystems. In this paper, a variable air volume (VAV) air-conditioning system is wholly analyzed with large-scale system theory based on “decomposition and coordination” strategy, and is partially controlled with iterative learning control (ILC) to improve the transient process. The dynamic and steady-state models were built for dynamic characteristic improvement and global optimal control, respectively. For the evaluation of the control policy, the physical tests were implemented on HVAC experimental platform. Results show that all control subsystems work coordinately with “decomposition and coordination” strategy under variable loads. The good performance of subsystems can be maintained under ILC strategy when working points change with the dynamic load.