神经网络解耦控制在中央空调系统中的应用  

The Application of Neural Network Decoupling Control in the Central Air-Conditioning System

作　　者：李界家[1] 陈浩[1] 李颖[2] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,辽宁沈阳110168 [2]沈阳仪表科学研究院,辽宁沈阳110043

出　　处：《沈阳建筑大学学报（自然科学版）》2012年第1期182-186,共5页Journal of Shenyang Jianzhu University：Natural Science

基　　金：国家自然科学基金项目(60874103)

摘　　要：目的针对中央空调控制系统中对温度的控制和湿度控制存在耦合问题,难于稳定运行的情况,提出神经网络解耦控制方案.方法通过引入神经网络环节,对多变量系统进行解耦,解耦后的子系统变为单变量系统.并通过matlab软件对其仿真.结果控制器满足了控制要求,并达到一定的精度,温湿度在8 min左右达到目标值,温度不随湿度的变化而变化,解耦成功.结论该控制器使得解耦后的多变量系统具备良好的动、静态特性,达到了理想的控制效果.In the central air-conditioning control system, the control of temperature and humidity has the cou- pling problem and it is difficult to operate stably. This paper presents neural network to decouple and control the system. The system is decoupled by using neural network sector. The decoupled subsystem is transformed into single variable system, and we simulate it by the matlab software. The controller can meet the control re- quirements, and achieve a certain accuracy, temperature and humidity in about 8min to reach target, the tem- perature does not vary with humidity changes, so the decoupling successes. This decoupled multivariate con- trol system has better static and dynamic performance and obtains the ideal control effect.

关 键 词：中央空调 神经网络 解耦控制 MATLAB仿真 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]