Multivariable Dynamic Matrix Control Method for Distillation Process Based on Subspace System Identification

The distillation process is characterized by strong coupling, multiple disturbances, nonlinearity, and other intrinsic complexities, rendering traditional control strategies inadequate in enhancing the control quality of distillation. This study leverages operational data from the methanol stripping tower in a coal-to-ethylene glycol distillation unit to employ a closed-loop subspace identification approach for modeling. Subsequently, a refined dynamic matrix control method is proposed and experimental investigations are conducted based on the established state-space model, ultimately achieving stable control of the distillation tower. The simulation results demonstrate the proposed subspace identification algorithm's ability to accurately predict the system's output. Furthermore, the dynamic matrix control method exhibits commendable resistance to disturbances and robustness when applied to the model, thereby ensuring effective control of the quality variables within the distillation tower.