Modeling and control of processes with output dynamic nonlinearity

The paper is concerned with the modeling and controlling of chemical processes with output dynamic nonlinearity, i.e., the process behavior is governed by a linear dynamics followed by a nonlinear unit with significant dynamics. Rather than modeling the overall process with a nonlinear model, it is proposed to represent the process by a composite model of a linear model (LM) and a feedforward neural network (FNN). The LM is to capture the linear dynamics, while the FNN is to approximate the remaining nonlinear dynamics. The controller, in correspondence, consists of two sub-controllers in a cascade fashion: a linear pole placement controller (PPC) designed based on the LM, and an iterative inversion controller (IIC) designed based on the FNN. Since the neural network is used to model the nonlinear component only, not the overall process, a relatively small size network is required, thus reducing computational requirement. The combination of linear and nonlinear control techniques results in a simple and effective controller for a class of nonlinear processes, as illustrated by the simulations in the paper.