Industrial SBR Process. Computer Simulation Study for On-line Estimation of Polymer Process Variables Using Neural Networks

MINARI, ROQUE J; STEGMAYER, GEORGINA S; GUGLIOTTA, LUIS M; CHIOTTI, OMAR A; VEGA, JORGE R

Mendoza (Argentina)

Congreso; 35 Jornadas Argentinas de Informatica e Investigación Operativa (JAIIO 06); 2006

An industrial emulsion copolymerization of styrene (S) and butadiene (B) carried out in a train of 7 continuous reactors operated under steady-state conditions is investigated. From “on-line” measurements of the reaction heat rate taken in all reactors of the train, a neural network (NN) model was developed for estimating several process variables, such as conversion, solids content, polymer production, particle diameter, and average copolymer composition. The NN model was trained and validated with a large set of simulation data generated through an available polymerization first-principles (FP) model, that was previously adjusted to several (steady-state and transient) experiments carried out in an industrial S-B rubber (SBR) process. Except for the copolymer composition, all variables can accurately be estimated through the proposed NN model. Due to the simplicity of the model and its reduced response time, the proposed NN is suitable for on-line estimation and closed-loop control purposes.