Experimental assessment of the progress variable space structure of premixed flames subjected to extreme turbulence

Abstract

Flamelet models of turbulent premixed combustion assume that (a) turbulent-transport and combustion processes can be decoupled and treated independently, suggesting that preheat and reaction zones remain layer-like and do not become highly fragmented and/or distributed. By further assuming (b) that the scalar-structure (e.g. the distribution of thermochemical quantities, like species mass fraction, vs. a control variable, such as a reaction progress variable) of a turbulent flame is akin to that in an associated laminar flame, detailed chemical properties can be introduce into a simulation at low computational cost via pretabulated flamelet tables derived from laminar flame calculations. The authors previously quantified conditions when assumption (a) remains valid. The present work assesses assumption (b) for turbulent Reynolds numbers that exceed those of previous studies by ∼ 7 × . Namely, planar laser-induced fluorescence (PLIF) of formaldehyde (CH2O), hydroxyl (OH), and methylidyne (CH) acquired jointly with Rayleigh scattering are used to produce joint PDFs (scatter plots) of the former vs. a progress variable (CR) derived from the latter. Regardless of the turbulence level the piloted Bunsen flames considered here were subjected to, peak-normalized conditional mean (CM) profiles obtained from such joint PDFs agree well with profiles derived from laminar flame calculations. Additionally, within the near field of modestly turbulent flames, two-term β-PDFs are found to accurately describe CR-distributions. However, agreement between β-PDFs and CR-distributions worsens with increased turbulence and axial distance from the burner. Small discrepancies between the measured CM profiles and those obtained from laminar calculations also emerged at highly turbulent conditions. For instance, under such conditions, laminar profiles of CH over predict the measurements where 0.6 < CR < 0.9. Nevertheless, the overall good qualitative agreement between the measured and simulated results supports the use of flamelet-based models for premixed flames subjected to extreme turbulence.