基于多模型切换的焦炉鼓风机调速系统智能控制策略研究及应用

来源期刊：冶金自动化2010年第3期

论文作者：张金花 张世峰 何福金

文章页码：10 - 13

关键词：焦炉； 鼓风机； 智能控制； 多模型

Key words：coke oven； blower； intelligent control； multiple models

摘    要：针对焦炉鼓风机系统具有工况复杂、非线性时变、大滞后及存在随机干扰等特点,提出了一种基于多模型切换的焦炉鼓风机调速系统的智能控制策略。把整个焦化工艺过程分两个不同工况:正常工况和强干扰状态(推焦、加煤等操作)下的非正常工况。从煤气发生量变化不同的角度又把正常工况分成两个工作区,通过采用基于最近邻聚类方法的RBF神经网络快速学习算法对这两个不同工作区分别建立被控对象的逆模型,并设计相应的两个逆模型控制器,模型之间是根据引入切换策略切换到相应控制器。系统在强干扰状态下切换到PI+Sm ith预估控制。仿真和应用结果表明,该多模型控制系统既保证了鼓风机系统的稳定性,又提高了对工况变动和干扰的适应性。

Abstract: For coke oven blower system with complex working condition,non-linear time varying,long delay and random interference,an intelligent control strategy for speed control system of coke oven blower based on multi-model switching is proposed.In whole coking process,there are two different working conditions:normal condition and non-normal condition under strong interference state(operation of pushing coke,charging coal etc.).Normal working condition is divided into two working regions in the view point of gas volume fluctuation.Through use of RBF neural network fast learning algorithm based on the nearest neighbor clustering method,inverse models of controlled objects were established for two different working regions respectively.Two coresponding inverse model controllers were designed.Model was switched to appropriate controller according to switching strategy.In strong interference condition,system was switched to PI+Smith predictive control.Application results show that the intelligent control system not only can ensure stability of blower system,but also improve adaptability to changes of working condition and interference.