HVAC Room Temperature Prediction Control Based on Neural Network Model

HVAC (Heating Ventilating &Air-conditioning) system is a nonlinear complex system with delay. It is very difficult to build a mathematical model of HVAC and implement model-based control. Since a BP (Back Propagation) neural network can fully approximate any nonlinear system with high recognition, a model of an HVAC system is built by using the BP neural network and such system is controlled by a generalized predictive control method. An experimental system of central air-conditioning is build in the lab and a simulation experiment shows that the neural network prediction model is robustness and it has a good tracking capability, the generalized predictive control has been chosen as an excellent controller of this nonlinear system.