Font Size: a A A

Control Design For VAV System Based On Neural Network Decoupling

Posted on:2012-11-28Degree:MasterType:Thesis
Country:ChinaCandidate:X Y LiFull Text:PDF
GTID:2132330338999441Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
How to saving energy and reducing pollution of air-conditioning are attracted more and more attention over the world, for the reason of energy crisis and pollution caused by enormous fuel burning. Facing the promotion of broadening sources and reducing expenditure, it is better to use optimum operation, which can facilitate the efficiency of devices and the system without having new energy resources. Thus, VAV air-conditioning system comes to widely used because of its comfort, energy saving, potential and flexibility. However, the actual use of VAV air-conditioning in China is worse than expection for many reasons, such as problems in construction or users'customs. In addition, for the multi-variable, strongly coupled, non-linear and time-variable of VAV system, the controlling, performance and management of VAV system are more difficult than CAV system.This article takes the actual VAV system project as research objects by collecting on-the-spot parameters. In the view of population situation that VAV system is hard to operate stable, the coupling characteristics of VAV system is deeply studied, and the control strategy of VAV system based on neural network decoupling is presented in this article.Firstly, this article discussed how to classify the VAV system after introducing the VAV system control methods and principle theory. Then, the article decides to study on the stability of VAV system by large reading and learning, which is the essential requirements of VAV system normal running. Next, according to VAV system characters, the mathematical model and parameters of VAV terminals are established in this article by means of experiments. In addition, the multiple variables decouping control strategy based on neural network is proposed on the base of traditional multiple variables decouping control methods. After completing the decoupling part, the neuron adaptive PID controller is also presented. And the initial weights are determined by use of traditional PID algorithm.Finally, the article introduces different VAV terminal systems in engineering projects. Additionally, a new way to realize the algorithm by using a language, which can be adapted by certain DDC is proposed and the language can be compiled in module, in which important parameters are send by interface. Certainly, the way should be discussed with VAV companies about its possibility.
Keywords/Search Tags:VAV system, decoupling control, neural network, PID
PDF Full Text Request
Related items