Hydroxyl time-series simulations for turbulent non-premixed flames

Picosend time-resolved laser-induced fluorescence (PITLIF) measurements for OH and CH in different turbulent diffusion flames show that nonreactive jet scaling is insufficient for describing minor-species time-series statistics. The axial growth rate of OH integral time scales is slower than those for velocity and mixture fraction (Z) in nonreacting jets, and the radial profiles for OH time scales are complex as compared to those for Z in nonreacting jets. A time-series simulation method was developed to study such observations for minor-species concentrations. It constructs time series for mixture fraction utilizing assumed parameters for the mixture fraction mean, rms, and power spectral density (PSD). For other scalars, time series are created by mapping the mixture fraction time series utilizing one-dimensional laminar-flamelet state relationships. Profiles for the mixture fraction inputs were taken from measurements in the same H₂/CH₄/N₂ flames for which OH and CH time-series data were available. OH mean and rms profile simulations were insensitive to variations in the input PSD for Z and the input OH state relationships. The mean rms profiles for mixture fraction affected strongly the predicted radial location of the OH peak concentration and the OH rms profile shape. The OH integral time scale prediction was unaffected by the mean Z, PSD for Z, or OH state relationships. The predicted radial profile for OH time scales captured most of the complex features unique to these scalars and were not observed in non-reacting jets. The simulation did not capture the measured variation in OH fluctuation rates with axial height or Reynolds number. Such differences resulted from uncertainties in the time scales for mixture fraction in reacting flows. Original is an abstract.