Lean fuel detection with nanosecond-gated laser-induced breakdown spectroscopy

Nanosecond-gated laser-induced breakdown spectroscopy (n-LIBS) has been used to quantify fuel mole fraction (ΧC2H4) in the cavity flameholder of a model high-speed combustor. The measurement locations selected, in the vicinity of the cavity shear layer, have low fuel concentrations. Previous n-LIBS measurements showed unexpectedly high values at these locations with expected low average fuel mole fraction (ΧC2H4) and intermittent fuel presence; thus, an effort was undertaken to understand potential sources of error with n-LIBS at low ΧC2H4. An improved direct spectrum matching (DSM) calibration matrix was thus constructed, focusing on low ΧC2H4 for both reacting and non-reacting conditions. Comparisons were made between n-LIBS measurements using two different laser systems for plasma generation, and the insensitivity of the signal to the specific laser was demonstrated for the first time. Uncertainty of the n-LIBS measurement technique at two different gas densities was analyzed through the processing of spectra acquired at known values of XC2H4. Measurements in the windtunnel, around the shear layer of the cavity flameholder, were conducted with a Mach-2 or Mach-3 freestream, and comparisons were made between n-LIBS results—using two DSM calibration matrices with different increments in XC2H4—and numerical results employing dynamic hybrid Reynolds-averaged Navier-Stokes and large-eddy simulation (DHRL). The high-resolution DSM matrix produced slightly lower XC2H4 than those from the coarse matrix, and comparisons with the simulation showed good agreement overall. For Mach-3 conditions, comparisons were made for both a non-reacting and a reacting cavity flameholder. Here, results showed a systematic difference between measured and simulated values of the mean XC2H4 within the shear layer but good (even excellent) agreement within the upper portion of the recirculation zone. For reacting condition, four reduced chemical kinetic models were implemented with the DHRL technique and compared to n-LIBS results. Overall, the four kinetic models produced a substantial range of results (e.g., temperatures and reaction progress), showing the importance of a validated diagnostic technique like n-LIBS for model testing.