基于模糊神经网络阶逆系统的发酵过程多变量解耦控制英文_.doc

Article ID: 1000?8152(2010)02?0188?05 Multivariable decoupling control based on fuzzy-neural network αth-order inverse system in fermentation process SUN Yu-kun, WANG Bo, JI Xiao-fu, HUANG Yong-hong (School of Electrical and Information Engineering, Jiangsu University, Zhenjiang Jiangsu 212013, China) Abstract: This paper proposes a nonlinear multivariable decoupling control strategy based on fuzzy-neural network αth-order inverse method that combines inverse system theory with fuzzy-neural network for fermentation process. A nonlinear inverse model is developed based on the reversibility analysis of the process model. A fuzzy-neural network αth-order inverse system is then constructed, which is cascaded with this process to transform the original nonlinear system to a pseudo-linear system. Finally, an expert controller is used to closed-loop synthesis. The effectiveness of the presented method is illustrated by a simulation experiment. Key words: bioprocesses; fuzzy-neural network; inverse system method; decoupling control; expert controller CLC number: TP273 Document code: A ??%‰???%α??fi?%Q???fi%????? ???, I ?, ???, ??? (??$? ????I???, ?? ?? 212013) ??: ??fi?{??%‰???%???, %$~???%‰???%α??fi?%Q???????{?. ???ffi????%??L, ??%‰???%%?Q???%????%?, ????fi%%‰??α??fi ??Q????‰??}???fi?, ?????}????????о?????. ????$?, %$%?? ??{?????Q???%?%???????%%%?, ?$??%???, t?f???fi?????{? ?????%???%???%%K%??%?A, %????, %???. }??: RKt???; %‰???%; ?fi?{?; ????; ?}??? 1 Introduction Bioprocess is a nonlinear multivariable coupling system for involving complex factors such as micro- bial cells growth, metabolism and so on[1] . Decou- pling control of this nonlinear multivariable system is a research topic of both theoretical and practical im- portance. Among these nonlinear system theories, the inverse system method is verified to be powerful[2,3] . Unfortunately, this method is based on an exact mathe- matical model of the plant, which is impossible to ob- tain in biop