Hongchao Dai

Whether reaction layers broaden at high Karlovitz numbers remains uncertain for premixed flames of heavy hydrocarbon fuels. This study examines the internal structure and turbulence evolution of prevaporized premixed n-decane/air jet flames at equivalence ratios of 0.5–1.4 and jet velocities of 100–300 m/s. Multi-species planar laser-induced fluorescence (PLIF) of OH/CH2O and CH/CH2O measurements with high spatial resolution were used to characterize the reaction-marker layers and CH2O-rich region, while simultaneous PIV/CH2O-PLIF measurements examined the local flow field across the flame. The flame marker layer thicknesses of highly turbulent flames were compared with laminar references obtained using the same image-processing procedure. Over nominal Ka values of 16–805, the OH-CH2O overlap layer thickness remained close to its laminar value, including for conditions located in the broken reaction zones regime of the Peters–Borghi diagram. CH measurements over a smaller Ka range likewise showed no systematic reaction-marker broadening. In contrast, the equivalent half-width of the CH2O-rich region reached approximately eight times its laminar value. The estimated Taylor microscale decreased toward the flame front, accompanying the broadening of the CH2O-rich region. These results show that, under the present conditions, substantial broadening of the CH2O-rich region can coexist with relatively thin reaction-marker layers beyond the conventional Ka = 100 regime boundary.

Quantitative species measurements are essential for optimizing hydrogen-fueled micro-mixing combustion, yet inherently weak spontaneous Raman scattering signals pose significant challenges for flame diagnostics. This study develops a systematic quantitative Raman spectroscopic methodology featuring simplified multi-pass excitation and automated spectral processing. Three-pass excitation using flat mirrors achieved three-fold signal enhancement with less than 7% energy loss, while avoiding air breakdown through afocal beam compression. An automated data processing workflow integrating Savitzky-Golay filtering, wavelet denoising, and Voigt profile fitting with a priori peak constraints was established to eliminate operator dependence and ensure reproducibility. Validation against standard McKenna flat-flame burner data demonstrates that the method achieves measurement deviations within 3% for major species (H2, O2, N2, H2O), with an expanded uncertainty of 2.8% (k = 1.96). These results reveal a nonlinear coupling between Raman signal intensity and equivalence ratio. As temperature rises, the reduction in molecular number density due to thermal expansion overwhelms the increase in cross-section, which leads to signal attenuation. The methodology was subsequently applied to measurements at 21 spatial locations in a newly designed rectangular micro-mixing hydrogen burner, capturing axial reactant gradients and flame front positions. This study establishes a reproducible quantitative Raman diagnostic methodology, providing a technical reference for species measurements in complex combustion environments.