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

In this paper, a neural network strategy for calculating the optimal set-po int trajectory in batch processes is implemented experimentally in the batc h synthesis of polymethyl-methacrylate (PMMA). It is shown that the optimal temperature trajectory which minimizes a desired performance index is a fu nction of the initial monomer concentration. This optimal trajectory is com puted on-line using a back propagation neural network. Both open-loop as we ll as closed-loop experiments are conducted in a 3 1 laboratory scale batch reactor and it is shown that a closed-loop scheme is necessary to keep the system on the desired trajectory. (C) 1999 Elsevier Science Ltd. All right s reserved.