Flame Surface Fractal Characteristics in Premixed Turbulent Combustion at High Turbulence Intensities

摘要

Because of the geometrical nature of the wrinkling of the flame surfaces in premixed turbulent combustion, the fractal approach is considered to be one of the most appropriate in power-law subgrid-scale models that are widely used in large eddy simulations. However, the available experimental data in literature on fractal characteristics of the premixed turbulent flame surfaces shows significant discrepancies among data sets, and the several different fractal algorithms used for evaluation may not produce consistent results. In this work, we analyzed premixed turbulent flame front images from an extensive experimental study using three different fractal algorithms and determined the fractal parameters: namely, fractal dimension D, inner cutoff ε_i, and outer cutoff ε_0. These images were obtained using laser-induced fluorescence of OH and Mie scattering on two Bunsen-type burners with diameters of 11.2 and 22.4 mm, with flames of propane-air with equivalence ratios of 0.8 and 1. Nondimensional turbulence intensity u'/S_L (ratio of fluctuating velocity to laminar flame velocity) was from 0.9 to 15, and the Reynolds number, based on the integral length scale, was from 40 to 467. The flame front surface areas were calculated for each set of results and compared with experimental turbulent burning velocities. The three fractal algorithms used for analysis in this work gave significantly different results for the same flame image sets. Further, nondimensional turbulent burning rates computed from the fractal parameters did not agree with the experimental turbulent burning data, except for those fractal results obtained by the box-counting method and only for u'/S_L ≤ 6. Implications of these results for flamelet models are discussed. In addition, the perimeter-ratio approach recently adopted to estimate the wrinkled flame surface area is discussed, and it is argued that the results obtained by the perimeter-ratio approach might not be unique.