On-line Estimation of Process Variables in an Industrial SBR Plant Using Neural Networks

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

Buenos Aires (Argentina)

Congreso; XXII Congreso Interamericano de Ingeniería Química - V Congreso Argentino de Ingeniería Química; 2006

This work investigates the production of Styrene-Butadiene Rubber (SBR) in an industrial continuous reactor train operated under steady-state conditions, with the aim of developing a soft-sensor for on-line monitoring several process and polymer quality variables. The sensor is based on two independent artificial neural networks (ANN); and allows estimating in each reactor of the train the following variables: the monomer conversion, the solids content, the polymer production, the particle diameter, the average copolymer composition, the average molecular weights, and the average branching degree. The required measurements are: (i) all the reagent flows feed into the first reactor; and (ii) the exchanged reaction heat in each reactor. The ANN models are trained and validated with a large set of simulation data generated through an available mathematical model, that was previously adjusted to several (steady-state and transient) experiments carried out in the industrial plant. The ANN models perfectly learn the training data set, and accurately predict independently-generated data (not included into the original training set). Due to the simplicity of the model and its reduced response time, the proposed ANN is suitable for on-line estimation and closed-loop control purposes.