Adaptive Control of Nonlinear Chemical Processes Using Neural Networks

Abstract An approach is proposed for nonlinear process modeling and control. Within the framework presented here, neural networks are employed for adaptively modeling nonlinear processes, on which model predictive control techniques are based. It is very difficult to provide a satisfactory representation of dynamic responses of the nonlinear processes over a wide operating region. A simple approach is proposed in which two components are used, the first component is the usual neural network which is trained to represent the steady state relationship, the second component serves as an adaptation mechanism to adaptively represent the dynamic response. The proposed approach has been demonstrated by a case study.