基于动态RBF-NN的温湿度解耦控制研究  

Research on Temperature and Humidity Decoupling Control Based on Dynamic RBF-NN

作　　者：孙冠琼 魏博 SUN Guan-qiong;WEI Bo(College of Chemical Engineering,Qinghai University,Xining 810000,China;Supervision and Control Center,Qinghai Western Hydropower Company,Minhe 810800,China)

机构地区：[1]青海大学化工学院,西宁810000 [2]青海西部水电有限公司监督计控中心,民和810800

出　　处：《自动化与仪表》2022年第12期24-27,33,共5页Automation & Instrumentation

基　　金：青海大学中青年科研基金项目(2019-QGY-14)。

摘　　要：温室控制系统变量较多,特别是温湿度控制间常存在强耦合现象,为解决这个问题,该文采用最近邻聚类算法对RBF-NN进行动态优化,提高对控制系统的在线辨识能力,再结合PID-NNC,建立基于动态RBF-NN的温湿度解耦控制策略。仿真结果表明,与常规的RBF-NN解耦控制方法相比,该策略解耦控制效果好,系统稳定性强,动态适应性好。There are many variables in the greenhouse control system,especially the strong coupling phenomenon between temperature and humidity control.In order to solve this problem,this paper uses the nearest neighbor clustering algorithm to dynamically optimize the RBF-NN to improve the online identification ability of the control system.Combined with PID-NNC,a decoupling control measurement strategy of temperature room temperature and humidity based on dynamic RBF-NN is established.The simulation results show that,compared with the conventional RBF neural network decoupling control method,this strategy has better decoupling control effect and strong adaptability.

关 键 词：温湿度控制 RBF-NN 解耦控制 最近邻聚类算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]