Optimal Control Framework for Alkane Separation in Semicontinuous Distillation

Abstract

Semicontinuous distillation (SCD) is a process intensification technique for separating N-component (N ≥ 3) mixtures. Its feature is forced circulation in three distinct operation modes: separating, discharging, and charging. To explore better control methods, a cascade model predictive control (MPC) strategy is proposed with a proportional-integral (PI) control loop, utilizing a multiple-input multiple-output transfer function (MIMOTF) matrix as the MPC prediction model. The MIMOTF matrix precisely represents the system’s dynamic behavior across the complete cycle. This method proves to be highly effective in controlling the entire SCD process while successfully meeting both product purity requirements and hydraulic constraints. A simulative economic comparison is conducted to assess the applicability of semicontinuous and conventional continuous systems in terms of the annual yield. The superiority of the cascade MPC strategy is demonstrated with remarkable reductions of 9.1% in separation cost (SC) and 8.9% in cycle time compared to the PI-only configuration.