基于逆系统的赖氨酸发酵多变量解耦内模控制  

Multivariable internal model decoupling control based on inverse system in lysine fermentation process

作　　者：孙晓天[1] 靳其兵[1] 张瑶[2] 徐海涛[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029 [2]江苏大学电气信息工程学院,江苏镇江212013

出　　处：《电子技术应用》2010年第2期91-94,共4页Application of Electronic Technique

基　　金：国家863资助项目(2008AA042131);国家973资助项目(2007CB714300)

摘　　要：针对赖氨酸发酵过程的时变、非线性和高耦合性,提出基于逆系统的赖氨酸发酵多变量解耦内模控制方法。根据动态递归模糊神经网络(DRFNN)的非线性辨识原理离线建立发酵过程的逆模型,将得到的逆模型串联在发酵系统之前,实现了发酵过程输入输出解耦线性化,从而得到伪线性系统;对复合后的伪线性系统采用内模控制。仿真结果表明,该方法能够适应赖氨酸发酵过程模型的不确定性和参数的时变性,具有较强的鲁棒性,且结构简单,易于实现。In view of lysine fermentative process time-variable, nonlinear and high coupling, multivariable internal model decoupling control based on inverse system in lysine fermentation process is proposed. First, according to the nonlinear identification theory of dynamic reeursive fuzzy neural network （DRFNN）,tbe nonlinear inverse model of the plant was built offline, the dynamic recursive fuzzy neural network （DRFNN） inverse model was cascaded before the fermentation system to be linear pseudo-linear system based on inverse system theory. Then using the internal model control to the pseudolinear system. The simulation example shows that this method can adapt lysine fermentative process model uncertainty and parameters of time-varying, has the strong robustness, and the structure is simple, easy to realize.

关 键 词：基于动态递归模糊神经网络(DRFNN) 逆系统 赖氨酸 解耦 内模控制 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]