Experimental and numerical analysis of stratified turbulent V-shaped flames

The present paper is devoted to (i) the experimental study of partially premixed combustion with strong equivalence ratio gradients, i.e., stratification of the reactive mixture and (ii) the numerical modeling of turbulent reactive flows in such situations where reactants are far from being ideally premixed. From a practical point of view, at least two variables are necessary to describe the local thermochemistry in this case: the mixture fraction ξ and the fuel mass fraction Y f are considered to represent respectively the local composition of the fresh mixture and the progress of chemical reactions. From the experimental point of view, the use of simultaneous imaging techniques allows the evaluation of both variables in terms of fuel mole fraction and temperature. In the present study, a combined acetone PLIF measurement and Rayleigh scattering technique is used. The influence of temperature on the fluorescence signal is corrected thanks to the knowledge of the local temperature through Rayleigh scattering measurements. Conversely, the influence of the acetone Rayleigh cross section can be evaluated with the local value of acetone mole fraction. Using the iterative procedure already described by Degardin et al. [Exp. Fluids 40 (2006) 452–463], the corrected fuel mole fraction and temperature fields can be obtained. Here the particular flow configuration under study is a stratified turbulent V-shaped flame of methane and air. In a first step of the analysis, the optical diagnostics are used to perform a detailed investigation of the flame thickness with a special emphasis on the influence of partially premixed conditions. In a second step, experimental data are used to evaluate the LW-P model as defined by Robin et al. [Combust. Sci. Technol. 178 (10–11) (2006) 1843–1870] to calculate turbulent reactive flows with partially premixed conditions based on an earlier analysis by Libby and Williams [Combust. Sci. Technol. 161 (2000) 351–390]. The closure problem raised by the mean scalar dissipation terms is also discussed in the light of experimental results. Since the usual closures for nonreactive flows are expected to be unsuitable to describe reactive scalar fluctuations decay a new modeling proposal based on the recent developments of Mura et al. [Combust. Flame 149 (2007) 217–224] is used. After a preliminary validation step where numerical predictions of the flame mean quantities are compared successfully with the experimental database, numerical simul