Distillation Column Machine Learning Model Controlled By Optimized DMC

In this paper, a successful implementation of ANN to build a realistic model from Aspen HYSYS simulation of the distillation column that separates LNG Hydrocarbons (C1 to n-C8) components is achieved. The machine learning model is used within the framework of the Model Based Predictive Control (MBPC) toolbox for optimum tuning; the control objective is to reduce variation of product purity due to operation constraints. This paper demonstrates how Aspen Hysys and Matlab-Simulink can be effectively used for control study. The results showed that the optimized DMC with NSGA-II control can provide offset-free dynamic control.