Multivariate Statistical Monitoring in an Industrial SBR Process. Part I: Prediction of Quality Variables

J.L. GODOY; R.J. MINARI; J.R. VEGA; J.L. MARCHETTI

Rosario, Argentina

Congreso; XIII Reunión de trabajo en Procesamiento de la Información y Control (RPIC 2009); 2009

UNR

This work investigates the production of Styrene-Butadiene Rubber (SBR) in an industrial continuous 7-reactor train operated under steady-state conditions. This work aims at developing a soft-sensor for on-line estimation of quality variables and statistical monitoring techniques capable of detecting and isolating physical sensor faults and non-measurable disturbances. In this first part, the proposed soft-sensor based on partial least squared (PLS) is found, validated and compared with a previous development based on neural networks. The soft-sensor was developed on the basis of dense multivariate sampling of steady-state (SS) simulations using a detailed model to on-line predicts the SS outputs. The achieved results demonstrate that the multivariate statistical approach provides a satisfactory sensitivity for anticipating changes in production and quality levels.