Prediction of clean octane numbers of catalytic reformed naphthas using FT-mir and PLS

An alternative method for predicting clean octane numbers (RON and MON) of catalytic reformed naphthas is examined. Using Fourier transform medium-infrared spectroscopy (FT-m.i.r.) and one-block and two-block partial least squares (PLS) analysis, improvements are obtained in sample throughput, delay time, precision (estimated repeatability and reproducibility) and environmental working conditions when compared with current standard methods. Smaller amounts of samples and reagents are needed. More than 300 samples were considered and monitoring samples were taken to verify model behaviour. Two types of reforming naphthas were encountered, depending on their aromatics content, and a different statistical model was developed for each of them. (C) 1997 Elsevier Science Ltd.