Control of a nonlinear pressure process in practice based on neural net models and exact-linearization

This paper deals with exact-linearization of nonlinear processes which can be modeled by feedforward neural networks. The authors propose two approaches. The first one uses nonlinear control theory to linearize a feedforward neural model of the involved process. The latter approach uses a feedforward neural network to exact-linearize the process directly. Both are used in the control of a pilot plant by using them in an IMC-structure. Both approaches gave comparable results