First imaging Fourier-transform spectral measurements of detonation in an internal combustion engine

Abstract

The Telops Hyper-Cam midwave (InSb 1.5-5.5 μm) imaging Fourier-transform spectrometer (IFTS) observed repeated detonations in an ethanol-powered internal combustion (IC) engine. The IC engine is a Megatech Corporation MEG 150 with a 1 in. bore, 4 in. stroke, and a compression ratio of 3 : 1. The IC combustion cylinder is made from sapphire permitting observation in the visible and infrared. From a distance of 3 m, the IFTS imaged the combustion cylinder on a 64 × 32 pixel array with each pixel covering a 0.1 × 0.1 cm 2 area. More than 14, 000 interferograms were collected at a rate of 16 Hz. The maximum optical path difference of the interferograms was 0.017 cm corresponding to an unapodized spectral resolution of 36 cm −1 . Engine speed was varied between 600-1200 RPM to de-correlate the observation time scale from the occurrence of detonations. A method is devised to process the ensemble of interferograms which takes advantage of the DC component so that the time history of the combustion spectrum can be recovered at each pixel location. Preliminary results of this analysis will be presented.

A Bayesian error model for quantifying methane and carbon dioxide column densities from hyperspectral measurements

“…As the hyperspectral mid-infrared spectral range includes absorption features from CO, CO 2 , H 2 O, SO 2 , NO, HCl, CH 4 , and other hydrocarbons (HCs), they have been used to detect and quantify the species densities in smokestacks [8,12,9], jet engine [13,7], automobile engine [14], and flare [15] gases. IFTSs generate a sequence of broadband images of the scene at very high framerates (e.g.…”

“…Additional uncertainties are introduced through measurement artifacts that cannot be removed entirely from the data. In the context of hyperspectral imaging of turbulent gas plumes, temporal variations in the scene intensity, caused by turbulent fluctuations in gas temperature and concentration as the interferogram is being formed, become convolved with intensity fluctuations, caused by wave interference, manifesting as broadband "scene change artifacts" (SCAs) in the recovered spectra [7,13]. Methods have been developed to remove these artifacts, and statistical descriptors have been derived to account for the related uncertainties [6], making their integration in a Bayesian framework possible.…”