Relationship between the smallest scale of flame wrinkles and turbulence characteristics of high-pressure, high-temperature turbulent premixed flames

Abstract

The mechanism determining the scale of the smallest flame wrinkles of high-pressure, high-temperature turbulent premised flames was investigated. The fractal inner cutoff of OH planar laser-induced fluorescence images, which is the smallest scale of flame wrinkles, was analyzed for burner-stabilized flames in a high-pressure chamber. Precise measurement of the energy spectrum of turbulence was also performed and the relationship between the intrinsic flame instability and flow turbulence was examined. Experiments were performed for CH4/air mixtures of 300 and 573 K at 0.1, 0.5, and 1.0 MPa. The experimental results clearly showed the Kolmogorov's similarity law for non-dimensional energy spectra of flow turbulence at high pressure and high temperature. A characteristic scale equivalent to the average vortex-tube diameter, l(v), which is about 10 times larger than the Kolmogorov scale revealed by recent direct numerical simulation, was used as a scale corresponding to the largest wave number of initial flow disturbances in an unburned mixture. At high pressure, the fractal inner cutoff, epsilon(i), and l(v) decrease with turbulence Reynolds number based on Taylor microscale, R-lambda, regardless of mixture temperature. The magnitude of epsilon(i) is close to l(v), when l(v) is larger than the characteristic instability scale corresponding to the maximum growth rate of flame instability, li. When R-lambda increases further and l(v) becomes smaller than l(v), epsilon(i) becomes almost constant. At atmospheric pressure, the relationship between l(v), l(v), and epsilon(i) was not obvious, but the correlation of epsilon(i) with the integral scale, l(g), was rather significant. These characteristics of epsilon(i) variation for high-pressure, high-temperature turbulent premixed flames can be explained using the scale-relation model based on l(g), l(v), and l(v) for R-lambda variation proposed in this study.