摘要

From the standpoint of polymer process applications based on near-infrared (NIR) spectroscopy and chemometrics, robust calibration models have been requested. Moreover, it has been expected to develop precise calibration models using sample data and spectra with small size because of savings in costs and labor. The purpose of the present paper is to describe some solutions for the above problems. An on-line monitoring system for molten polymers based on a Fourier-transform near-infrared (FT-NIR) spectrometer was developed. The influences of a change in the optical Fiber temperature and sample temperature have been investigated. Furthermore, a modified multiplicative scatter correction was proposed as a traceable compensation method for changing the effective pathlength by a change of the sample temperature. NIR spectra of linear low-density polyethylene (LLDEP), random polypropylene (RPP), block polypropylene (BPP), and ethylene vinyl acetate copolymer (EVA) in the melt and solid-states were measured by the on-line and laboratory FT-NIR system, and compared in detail. The sample sets for developing prediction models for the density of LLDPE were collected for three months at a plant, and the density of LLDPE was continuously monitored on-line for another three months using the calibration model. A prediction of the ethylene content in the melt states RPP and BPP was studied. The method using a difference spectrum between a RPP and a BPP spectrum has been proposed to transfer from the BPP calibration set to the RPP calibration set. The calibration models for the vinylacetate (VA) concentration in EVA copolymers has been investigated and proposed to make use of band shifts associated with concentration changes in the VA for improving of the models. Finally, a practical calibration correction method for an on-line monitoring system has been studied. It has been demonstrated that the proposed method, considering the wavenumber shift and path-length correction, is useful for a calibration correction in a real plant.

  • 出版日期2010-5