Model Identification and Controller Simulation for Distillation Column

For any complex industrial process, the control system simulation requires the mathematical model in any form of the plant being controlled. If the plant model is not available then control system simulation is too challenging using conventional techniques such as PID tuning methods. In this paper, the input-output data collected pertaining to an industrial distillation column is used to identify various models (eg. FOPDT, TF, SS, ARX, ARMAX) of the plant using system identification toolbox of MATLAB. The best fit model is then selected for the control system simulation. PID and MPC controllers are designed for the best fit model, and their performance has been compared.