Generalized Predictive Control of HVAC Systems Based on Fuzzy RBF Networks

The paper presents a predictive control scheme of heating ventilating air conditioning (HVAC) systems based on fuzzy RBF networks. The input space is divided by means of clustering, and the HVAC prediction model is set up using RBF networks. The generalized predictive control of the system is achieved by a fuzzy RBF controller. Practical experiments have demonstrated that the proposed control system is capable of timely and accurate tracking of the temperature set point in case of variations in refrigerated water temperature, blast pressure, and humidity, etc. as well as environment disturbances. As a result energy consumption is reduced and human comfort is assured