Design of Integrated Real –time Optimization and Adaptive Model Predictive Control for Distillation Column

This paper presents two integration approaches of real-time optimization (RTO) and adaptive model predictive control (MPC) in the 2-layer and 3-layer structures and their applications to the distillation column. The 2-layer structure is constructed based on the conventional hierarchical control layer. In the 3-layer structure, the quadratic programming (QP) layer is inserted to compute the feasible set points for the MPC layer. The uncertainty in the process model is tackled by using the adaptive MPC with the switching control methodology. Each operating condition of the process is denoted by a corresponding model. The application to distillation column illustrates the differences in the transient responses of two integration structures. The simulation result shows that the 3layer structure gives stability region larger than that of the 2-layer structure in enforcing the stability of the system.