Intelligent control scheme of a batch process

This paper describes an intelligent control scheme for the reaction stage of a batch polypropylene reactor. With time varying and nonlinear characteristics, it is quite difficult to control a batch unit only using traditional control methods. In this paper, bang-bang control, fuzzy control and traditional PID control are combined together to form an intelligent compound control system, based on the actual technology and the rich experience of skilled operators. A multi-layer feed-forward neural network is used to model the process. The efficiency of the proposed control scheme is illustrated. It solves the conflict between steady state error and dynamic responsiveness.