Mid-IR hyperspectral imaging of laminar flames for 2-D scalar values.

Abstract

This work presents a new emission-based measurement which permits quantification of two-dimensional scalar distributions in laminar flames. A Michelson-based Fourier-transform spectrometer coupled to a mid-infrared camera (1.5 μm to 5.5 μm) obtained 256 × 128pixel hyperspectral flame images at high spectral (δν̃ = 0.75cm(−1)) and spatial (0.52 mm) resolutions. The measurements revealed line and band emission from H2O, CO2, and CO. Measurements were collected from a well-characterized partially-premixed ethylene (C2H4) flame produced on a Hencken burner at equivalence ratios, Φ, of 0.8, 0.9, 1.1, and 1.3. After describing the instrument and novel calibration methodology, analysis of the flames is presented. A single-layer, line-by-line radiative transfer model is used to retrieve path-averaged temperature, H2O, CO2 and CO column densities from emission spectra between 2.3 μm to 5.1 μm. The radiative transfer model uses line intensities from the latest HITEMP and CDSD-4000 spectroscopic databases. For the Φ = 1.1 flame, the spectrally estimated temperature for a single pixel 10 mm above burner center was T = (2318 ± 19)K, and agrees favorably with recently reported laser absorption measurements, T = (2348 ± 115)K, and a NASA CEA equilibrium calculation, T = 2389K. Near the base of the flame, absolute concentrations can be estimated, and H2O, CO2, and CO concentrations of (12.5 ± 1.7) %, (10.1 ± 1.0) %, and (3.8 ± 0.3) %, respectively, compared favorably with the corresponding CEA values of 12.8%, 9.9% and 4.1%. Spectrally-estimated temperatures and concentrations at the other equivalence ratios were in similar agreement with measurements and equilibrium calculations. 2-D temperature and species column density maps underscore the Φ-dependent chemical composition of the flames. The reported uncertainties are 95% confidence intervals and include both statistical fit errors and the propagation of systematic calibration errors using a Monte Carlo approach. Systematic errors could warrant a factor of two increase in reported uncertainties. This work helps to establish IFTS as a valuable combustion diagnostic tool.