Neural network control for an intelligent air handler in an air-conditioning system

Abstract

Many commercial air-conditioning systems in hot and humid areas like Singapore are operated throughout the year. There are two main classifications for these systems: the constant air volume (CAV) and variable air volume (VAV) systems. Currently, both CAV and VAV systems are designed to control the indoor temperature only, while leaving the space relative humidity to float during part load operation. Under varying load and external environmental conditions, general proportional-integral-derivative (PID) control technique may not be appropriate for the regulation of both the temperature and the relative humidity.