Near-Infrared Spectroscopy for On-Line Monitoring of Lube Base Oil Processes

Abstract

Near-infrared (NIR) spectroscopy has been utilized for monitoring the pour point of the lube base oil (LBO) process. Using partial least-squares (PLS) regression, necessary spectral features were successfully extracted and correlated to the reference pour-point data. In a preliminary laboratory feasibility study, it was found that PLS calibration performance was largely governed by the accuracy of the reference pour-point analysis. The same NIR spectroscopic methodology was moved to monitor the pour point continuously in an on-line manner. At this time, the NIR spectrometer was calibrated with the existing on-line pour-point analyzer, which was more accurate. In comparison with the on-line pour-point analyzer, NIR spectroscopy provided faster and more repeatable analytical data. With the help of NIR analysis, LBO process variations during product switch can be detected earlier, and necessary process-control strategy can be applied to improve the process efficiency.