A fuzzy decision-making approach to temperature control in air-conditioning systems

The paper describes a fuzzy model-based predictive control scheme for controlling the supply air temperature in an air-conditioning system. The control objective is defined in terms of a fuzzy goal and a fuzzy decision-maker is used to find the membership function of the optimum fuzzy control signal at each sample time. Fuzzy predictions of the future behaviour of the system are generated using a generic neurofuzzy model. The neurofuzzy model is identified using training data obtained from computer simulations of cooling coils of the same type as that used in the air-conditioning system. A conditional defuzzification scheme is used in order to reduce the control activity. The performance of the fuzzy model-based controller is compared to that of a conventional controller on a laboratory test rig. Results are presented that show that satisfactory control of the supply air temperature is possible, at both high and low air flow rates, with a minimum of control activity.