On the use of extended-wavelength FTIR spectra for the prediction of combustion properties of jet fuels and their constituent species

Abstract

A Fourier transform infrared (FTIR) spectra-based strategy was developed for estimating three important combustion properties of jet fuels and their constituent hydrocarbon species: ignition delay time (IDT), net heat of combustion (NHC), and derived cetane number (DCN). This approach leverages the strong sensitivity of combustion behavior to the different infrared absorption features of hydrocarbon fuels. Gas-phase FTIR spectra of pure hydrocarbons and jet fuels in the 2–15.38 μm wavelength range were used to train elastic-net regularized linear models, with the model parameters optimized separately for each property. The results from these models were compared with the results from previous models, which utilized only a limited spectral region in the wavelength range 3.3–3.55 μm. The new, optimized models developed in this work show significant improvement in predictive performance compared to the previous models for all three properties. The models’ prediction errors on test data (blends of conventional and alternative jet fuels) are found to be comparable to the standard property measurement uncertainties, demonstrating the value of infrared spectral analysis as a low-volume prescreening tool for accurate property estimation of both conventional and alternative jet fuels.