Neural Network Model Predictive Control of Non- Linear Biopolymerization Process

This paper focuses on the developing a mechanistic model of the biopolymerization process and linking with the feedforward neural network model to obtain a hybrid model of lipase-catalyzed ring-opening polymerization of ε-caprolactone (ε-CL) for Poly (ε-caprolactone) PCL production using feed forward neural networks and control using model predictive control. Model predictive control (MPC) has been applied to predict the molecular weight of the biopolymer and consequently control the molecular weight of biopolymer. The result shows that MPC is able to track reference trajectory and give optimum movement of manipulated variable.