Experimental implementation of an on-line optimization scheme to batch PMMA synthesis

Abstract

In this paper, a neural network strategy for calculating the optimal set-point trajectory in batch processes is implemented experimentally in the batch synthesis of polymethyl-methacrylate (PMMA). It is shown that the optimal temperature trajectory which minimizes a desired performance index is a function of the initial monomer concentration. This optimal trajectory is computed on-line using a back propagation neural network. Both open-loop as well as closed-loop experiments are conducted in a 3 l laboratory scale batch reactor and it is shown that a closed-loop scheme is necessary to keep the system on the desired trajectory.