Improved kinetic data from analysis of complex hydrocarbon mixtures by using SIMCA

A SIMCA model was developed for classification of C10-bicycloalkanes to assist evaluation of kinetic experiments involving complex mixture of C10-hydrocarbons. The influence of different pre-processing methods, such as square root and cubic root scaling, and transformation methods, including autocorrelation transform and spectral features, on the classification of C10-bicycloalkanes was investigated. While the transformation of mass spectra into spectral features was found essential for modeling of spectra of C10-bicycloalkanes, the intensities scaling was only of little importance for the quality of the SIMCA models. The best SIMCA model was successfully applied in detailed analysis of C10-bicycloalkanes originating from decalin isomerization and contributed to identification of the decalin isomerization mechanism.