PREDICTION OF MOONEY VISCOSITY OF A STYRENE - BUTADIENE RUBBER THROUGH A HYBRID POLYMERIZATION MODEL

GERARDO MARTINEZ DELFA; ROQUE MINARI; CARLOS EUGENIO BOSCHETTI; JORGE VEGA; LUIS GUGLIOTTA

Bahía Blanca

Simposio; IX Simposio Argentino de Polímeros; 2011

UNS-CONICET

Industrial styrene-butadiene rubber (SBR) is mainly produced through emulsion copolymerization processes. Based on the rubber end use (i.e. tires, seals, soles), the fit of the SBR quality properties to given specifications is a key task at the production plant, in order to guarantee an adequate processability of the final rubber.Typical SBR processability parameters include Mooney viscosity (MV), delta Mooney, and black incorporation time, which are directly measured in laboratory on the final rubber. Usually, MV is the most accepted quality variable (Baranwald and Stephens, 2001); but it is measured with typical delays of 2 hours, and therefore inadequate for implementing on-line control strategies. Several polymerization processes are controlled through open-loop strategies derived on the basis of first-principle (FP) models of the polymerization. As far as the authors are aware, no FP model able to predict the rubber MV has been published yet. Alternatively, some artificial neural networks (ANN) models have been recently proposed to estimate MV (Padmavathi et al., 2005; Martinez Delfa et al., 2009).Hybrid models emerge as an attractive alternative to overcome the problem of rubber MV estimation. A hybrid model combines the available knowledge on the process, incorporated in the FP model, with an ANN used for estimating the unavailable and non-modeled process variables (e.g., the MV). In this work, the SBR emulsion copolymerization is investigated with the aim of developing a hybrid mathematical model for predicting the gravimetric monomer conversion, the molecular weight distribution (MWD) of the copolymer, and the MV of the SBR.