Global optimal control of variable air volume air-conditioning system with iterative learning: an experimental case study

Abstract

The air-conditioning system in a large commercial or high-rise building is a complex multi-variable system influenced by many factors. The energy saving potential from the optimal operation and control of heating, ventilating, and air-conditioning (HVAC) systems can be large, even when they are properly designed. The ultimate goal of optimization is to use the minimum amount of energy needed to improve system efficiency while meeting comfort requirements. In this study, a multi-zone variable air volume (VAV) and variable water volume (VWV) air-conditioning system is developed. The steady state modes and dynamic models of the HVAC subsystems are constructed. Optimal control based on large scale system theory for system-level energy-saving of HVAC is introduced. Control strategies such as proportional-integral-derivative (PID) controller (gearshift integral PID and self-tuning PID) and iterative learning control (ILC) are studied in the platform to improve the dynamic characteristics. The system performance is improved. An 18.2% energy saving is achieved with the integration of ILC and sequential quadratic programming based on a steady-state hierarchical optimization control scheme.