Tomographic Imaging of Flames: Assessment of Reconstruction Error Based on Simulated Results

Temperature and concentration distributions of a horizontal section through a simulated stationary test flame were reconstructed with the help of computer tomography. The study uses temperature and concentration distributions equal to published time-averaged data for the Sandia D flame. Projection values are obtained by simulated line-of-sight integration for two water vapor (H2O) absorption lines in the 6930-6940 cm-1 range, where direct absorption spectroscopy and wavelength modulation spectroscopy with second harmonic frequency detection are considered. The projected values obtained from line-of-sight measurements are deconvoluted by tomographic methods, where Abel inversion and filtered backprojection are investigated. The tomographic reconstruction yields spatially resolved spectroscopic data, which is then used to calculate distributions for temperature and concentrations. Results illustrate that tunable diode laser absorption spectroscopy artifacts are caused by relatively small errors in reconstructed spectroscopy data. In regions with low-signal magnitudes, tomographic errors are amplified when temperature and concentrations are calculated using two-line ratio thermometry. Although this study indicates that the selection of absorption lines has an impact on the quality of tomographic reconstructions, results nevertheless show that second harmonic frequency detection, in conjunction with tomography, is a viable method to reconstruct spatially resolved temperature and concentration distributions for stationary flames.