Applicability of high-absorbance MIR spectroscopy in industrial quality control of reformed gasolines

Partial least squares (PLS), polynomial partial least squares (polynomial-P LS), locally weighted regression (LWR) and genetic inside neural network (G INN) algorithms were used to develop models for predicting motor octane num ber (MON) from non-leaded and catalytically reformed gasolines. Medium infr ared (mid-infrared) spectra were obtained on liquid samples and chemometric ally processed in order to get acceptable predictive models which allow the ir use for routine industrial quality monitoring. As MIR spectra currently present peaks with high absorbances, the presence and influence of nonlinea rities was sought comparing the broadly-used PLS method with several other algorithms specially designed to cope with such influences (polynomial-PLS, local regression and neural networks). Their prediction abilities; i.e., s tability and global prediction error when predicting new samples as well as their usefulness for routine industrial control were studied. (C) 1999 Els evier Science B.V, All rights reserved.