Title :
Neural net based model predictive control of a chaotic continuous solution polymerization reactor
Author :
De Souza, Maurfcio B., Jr. ; Pinto, Jose Carlos ; Lima, ENRIQUE L.
Author_Institution :
UFRRJ, Rio de Janeiro, Brazil
Abstract :
Continuous polymerization processes may present extreme sensitivity to small changes in the operational conditions. For example, a CSTR for the homopolymerization of VA (vinyl acetate) showed multiple steady-states and oscillatory behavior. In this paper, a predictive control formulation that uses an ANN (artificial neural network) as the internal model of the process is employed to stabilize this reactor and make it less sensitive to disturbances.
Keywords :
chemical technology; neurocontrollers; polymerisation; predictive control; process control; CSTR; artificial neural network; chaotic continuous solution polymerization reactor; homopolymerization; internal model; multiple steady-states; neural net based model predictive control; operational conditions; oscillatory behavior; vinyl acetate; Chaos; Continuous-stirred tank reactor; Inductors; Industrial control; Neural networks; Polymers; Predictive control; Predictive models; Steady-state; Temperature control;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.716998
