A simple and compact form of reduced-order distillation model especially suitable for real-time applications is proposed. For this purpose, a modular collocation approach with the cubic spline interpolation function is developed and applied to an underlying distillation model which is constructed based on the McCabe and Thiele assumptions plus constant tray holdups.
To evaluate the performance of the model, numerical simulations are carried out for the case of dynamics as well as steady states. As a consequence, it is found that the proposed reduced-order model gives better approximation than those obtained by the conventional reduced-order model with the Lagrange interpolation function.