球磨煤机隶属函数型模糊神经网络解耦控制方法  

Research on fuzzy neural network decoupling control method based on membership function for ball mill

作　　者：刘鑫屏[1] 刘吉臻[1] 楼冠男[1] 平玉环[2] 

机构地区：[1]华北电力大学控制科学与工程学院,河北保定071003 [2]华北电力大学科技学院,河北保定071003

出　　处：《中国电力》2010年第4期72-75,共4页Electric Power

基　　金：国家863重点资助项目(2007AA04Z163)

摘　　要：设计适合于火电厂球磨煤机的解耦控制方法,有利于提高热工自动化水平。依据模糊控制技术和跟踪解耦的思想并结合神经网络控制,介绍了一种不依赖于被控对象精确数学模型的隶属函数型模糊神经网络多变量解耦控制方案。将该种方法应用于球磨煤机解耦控制系统,对耦合强烈的磨煤机出口温度、入口负压子系统进行解耦,获得期望的多环结构。对解耦结果进行了仿真研究。对比试验显示该解耦方法具有满意的动态特性和较强的鲁棒性。It is good for improving automation level to design decoupling method suitable for hall mill. According to fuzzy control technology and tracking deeoupling idea, combined with neural network control, a fuzzy neural network decoupling control method based on membership function, which is independent of precise mathematic model, was introduced and was applied in ball mill control system. The subsystem of outlet temperature and inlet negative pressure with strong coupling was decoupled to get multi-loop structure. Simulation example for the design is done. The results show that the deeoupling algorithm has satisfactory dynamic behaviors and strong robustness.

关 键 词：球磨煤机制粉系统 模糊控制技术 神经网络技术 隶属函数 多变量解耦 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]