Robust neural network controller for variable airflow volume system
The HVAC industry expects the automation and control systems to perform well throughout the year without the requirement to re-tune the system and to conserve energy. To satisfy these requirements it is necessary to provide a basic control algorithm that will respond well to the presence of nonlinear behaviour in HVAC equipment. The PID algorithm has to be enhanced to handle the highly nonlinear functionality, range of operation and robustness. The neural network is one of the best candidates to deal with these issues. However, it is important to address stability and disturbance properly, to obtain optimal performance of the neural control system. The design and application of a robust neural network algorithm are discussed and how it compliments the fixed proportional control algorithm to provide the desired functionality as well as the adaptation of the VAV control system for a wide range of disturbances and parameter changes.
