Topology and burning rates of turbulent, lean, H2/air flames

• Published in: Combustion and flame Vol. 162; no. 12; pp. 4553 - 4565
• Main Authors: Amato, Alberto, Day, Marcus, Cheng, Robert K., Bell, John, Dasgupta, Debolina, Lieuwen, Tim
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01-12-2015; Elsevier
• Subjects: Flame stretch; INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; Leading points; Premixed flames; Turbulent combustion; Flame stretch; Turbulent combustion; Premixed flames; Leading points
• ISSN: 0010-2180; 1556-2921
• DOI: 10.1016/j.combustflame.2015.09.010

Improved understanding of turbulent flames characterized by negative consumption speed-based Markstein lengths is necessary to develop better models for turbulent lean combustion of high hydrogen content fuels. In this paper we investigate the topology and burning rates of turbulent, lean (ϕ = 0.31), H2/air flames obtained from a recently published DNS database (Aspden et al., 2011). We calculate local flame front curvatures, strain rates, thicknesses, and burning velocities and compare these values to reference quantities obtained from stretched laminar flames computed numerically in three model geometrical configurations—a counterflow twin flame, a tubular counterflow flame and an expanding cylindrical flame. We compare and contrast the DNS with these model laminar flame calculations, and show both where they closely correlate with each other, as well as where they do not. These results in the latter case are shown to result from non-flamelet behaviors, unsteady effects, and curvature-strain correlations. These insights are derived from comparisons conditioned on different topological features, such as portions of the flame front with a spherical/cylindrical shape, the leading edge of the flame, and portions of the flame front with low mean curvature. We also show that reference time scales vary appreciably over the flame, and characterizing the relative values of fluid mechanic and kinetic time scales by a single value leads to erroneous conclusions. For example, there is a two order of magnitude decrease in chemical time scales at the leading edge of the front relative to its unstretched value. For this reason, the leading edge of the front quite closely tracks quasi-steady calculations, even in the lowest Damköhler number case, DaF∼0.005.