Two Case Studies for Applying Model Predictive Controllers on Chemical Processes

Implementation of model predictive control as the most famous advanced process control method, in real processes has some practical issues that are ignored in many simulation studies and needs more attention especially in implementing the controller. For this purpose, in this paper two chemical processes are simulated in HYSYS software as a more realistic environment that exhibits many properties of real plants. The control and identification is performed by connecting HYSYS and MATLAB softwares in a real-time manner. Some of the main practical issues that are considered here are: pre-test of the plant, test design and identification requirements, dealing with low-level control loops and controller tuning to attain the best performance. Simulation results show that without having enough knowledge on such important points, practical implementation of this very useful advanced process control method will be very challenging.