PI-IMC-Based Controller Design for an Ethanol–Water Distillation Column

Abstract

Distillation column is an important unit operation for multicomponent separation. However, its highly complex and nonlinear characteristics require an effective control strategy to maintain consistent product purity. This work focuses on the design and performance evaluation of an IMC-PI controller for a lab-scale distillation column. A First-Order Plus Dead-Time (FOPDT) model is developed based on experimental data, with model accuracy evaluated using R² and RMSE values of 0.8857 and 0.0119, respectively. Three PI controller tuning methods; AMIGO, Cohen-Coon, and IMC are applied and assessed based on setpoint and disturbance rejection response using Scilab XCOS simulation. The results show that IMC provides the best balance between overshoot and settling time, with no oscillations occurring. Scilab XCOS simulations are used for tuning and robustness analysis of the IMC PI-controller by introducing variations of the controller time constant,