Feed-forward Decoupling And Control Basing Adaptive Linear Neural For The Vav Box Side Of Variable Air Volume Air-Conditioning System

Of all the Building Automation Systems (BAS), heating, ventilating and air-conditioning (HVAC) systems are most important from the viewpoint of saving energy and improving the quality of the indoor environment. Therefore, adopting efficient control strategies for air-conditioning plays a vital role in developing improved energy management control (EMC) systems for intelligent buildings (IB). Nowadays, VAV air-conditioning system has gradually become most popular in China and abroad because of its significant energy saving. However, since VAV system is multivariable, strongly coupled , nonlinear and time variant, its design, performance and management are more difficult than CAV system's. It's an important reason that limits the vav system for a greater use. The paper takes the vav box as the object for studying. It has analyzed the coupling relation between the control circuits of the vav box side in details. And also, the paper has erected a model for the control by the means of experiment. The concept of adaptive linear neural net control for multi-variables system has been applied in this paper. According to the vav box side of vav system, this paper has concluded a plan of adaptive linear neural net decoupling and control. In fact, the scheme is the concept of feed-forward decoupling and control. In order to prove the plan's effectiveness, the paper has done jobs in two aspects. One is emulation via the tool of Matlab, the other is doing experiments in real system. Both of the trying are pleased. And they have proved the plan's validity. Also, the paper has provided a program for the system's decoupling and control. It is designed by the tool of visual basic. The program exchanges data with the system of Metasy in the mode of VB-DDE-METASYS. And it has fine operating interface.