基于集成学习算法的生物质气化炉改进模糊双闭环控制系统  

Improved fuzzy dual closed-loop control system for biomass gasifier based on integrated learning algorithm

作　　者：刘辉[1] 庞佑霞[2] 唐勇[2] 

机构地区：[1]长沙学院电子与通信工程系,长沙410003 [2]长沙学院机电工程系,长沙410003

出　　处：《仪器仪表学报》2013年第3期643-649,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(50975032);湖南省"十二五"重点建设学科资助项目

摘　　要：生物质气化过程是具有弱随机性的物理化学过程,针对其工艺特点本文提出一种基于集成预测和改进模糊的控制算法。外环采用基于SVM的AdaBoost集成学习方法,预测温度变化趋势;内环采用了遗传模糊算法,将预测结果和含氧量反馈值作为输入,在稳定炉顶层温度和降低含氧量这对矛盾之间,寻求最优鼓风机的输出,对一次风进风量进行调节。仿真和现场运行数据证明了方法够满足气化炉的工艺要求和控制目标。The biomass gasification process is a physical and chemical process with weak randomness. This paper presents a control algorithm based on integrated forecasting and improved fuzzy control aiming at its technological characteristics. AdaBoost ensemble learning method based on SVM is adopted to predict the temperature changing trend in the outer loop;genetic fuzzy algorithm is adopted in the inner loop,and the predicted result and the oxygen content feedback value are taken as the inputs;through compromising stabilizing furnace top temperature and reducing oxygen level, the optimal blower output is sought, and the air flow of primary air is adjusted. Simulation and field operating data prove that the proposed method can meet the gasifier production technological requirement and the control obiective.

关 键 词：气化炉 支持向量机 ADABOOST 模糊控制器 遗传算法 

分 类 号：TP273.21[自动化与计算机技术—检测技术与自动化装置] TH79[自动化与计算机技术—控制科学与工程]