The Control Research Of VAV Based On The Predictive Control Technique

The Energy-consumption of HVAC has been a relative part of the sum building energy-consumption in modern society. To save energy and improve the indoor air quality, the automatic control system of HVAC has been the most important part of the building automation, and it is significant to adopt effective air-conditioning method to enhance the energy management of intelligent architecture. Nowadays VAV air-conditioning system has been the most popular air-conditioning system home and abroad for its considerable energy-saving potential. However, since VAV system is multivariable, strongly coupled, nonlinear and time variant, its design running and management are more difficult than that of CAV. It restricts its application and popularization badly in our country.The paper took control strategy of VAV air-conditioning system as cut-in, and applied the NN(neural network) technique which has been wide used in all kind of fields and the predictive control technique which has grown from industrial process control to control design of the VAV system. For its characteristic as multivariable, strongly coupled, non linear and time variant, if the control method is still a single variable method, the effect is not satisfying. So it is necessary to decouple the VAV air-conditioning system before designing the control loop. At the beginning of the paper, neural network technique was used to design the decoupling loop of the VAV air-conditioning system. Then by using latent generalized predictive control technique I designed the control system for the decoupled object. The paper took a VAV system as example, integrated the two techniques to work out the neural network decoupler and the predictive controller to complete the track control of the object. It followed that the control effect accorded with the initial assumption.