SciELO - Scientific Electronic Library Online

 
vol.72 issue147POLY-ANILINE SYNTHESIS BY ELECTROCHEMICAL OXIDATIONNEUROSCHEME: A MODELING LANGUAGE FOR ARTIFICIAL NEURAL NETWORKS author indexsubject indexarticles search
Home Pagealphabetic serial listing  

DYNA

Print version ISSN 0012-7353

Abstract

HERNANDEZ G., HUGO  and  AGUIRRE C., JAIME. MOLECULAR WEIGHT DISTRIBUTION ESTIMATION IN BATCH POLYESTERIFICATION USING A MODEL-BASED SOFT SENSOR. Dyna rev.fac.nac.minas [online]. 2005, vol.72, n.147, pp. 65-73. ISSN 0012-7353.

In this paper, a methodology for the development of a first principles model based soft sensor is proposed for the estimation of quality properties of chemical products. A practical example is presented where a soft sensor is developed to estimate the molecular weight distribution of a polymer during the polyesterification of a carboxylic diacid and a diol (A-A/B-B) in a stirred-tank batch reactor. Results obtained in the validation stage show that it is possible to employ a model-based soft sensor as the input element of an automatic quality control system. The estimation error of the soft sensor is comparable to the measurement error of physical sensors currently available, but the former has the advantage of providing on-line and real-time measurements.

Keywords : Soft sensor; Quality; Polymers; Molecular weight distribution; Process control.

        · abstract in Spanish     · text in Spanish     · pdf in Spanish