Hydrodynamic instability and shear layer effects in turbulent premixed combustion

Abstract

A turbulent premixed plane jet flame is analyzed by large-eddy simulations. The analysis shows that the flame front wrinkling is strongly influenced by the shear layer effect when the gas expansion effects are small leading to larger flame front amplitudes at the flame base than at high gas expansion ratios. However, the hydrodynamic instability effect induces a continuously increasing flame front amplitude which yields an enhanced flame pocket generation at the flame tip. Both phenomena influence the magnitude of the turbulent burning area and burning area rate response through the flame front deflections which are determined by the contribution coefficient. This coefficient represents the mutual interaction between the flame and the flow. At low gas expansion ratios, the total heat release rate spectra of the turbulent flame are wider in terms of dominant modes at Strouhal numbers which are linked to the mean flame height oscillations. Thus, at low gas expansion ratios, the vortex-flame interaction is less damped by the flame in the sense that vortices can perturb the flame front stronger. The total heat release rate trend of St−2.2 previously found for a round jet flame is also determined for the current slot jet at realistic gas expansion ratios indicating a general tendency to transfer energy from large to small flame structures. At high gas expansion ratios, an increasing Markstein length leads to an energy transfer between neighboring dominant modes in the low frequency range 1 < St < 10 and the burning area rate response becomes more important for the total heat release rate spectra of the turbulent slot flames which agrees with recent findings for a laminar premixed plane flame.

Dynamic responses of anchored ducted flames to harmonic velocity forcing

“…The discrete vortex method has been successfully applied in previous studies to simulate the vortex dynamics and flow fields in bluff body flows. 31,32 In the current model, we assume that the vortex generation only occurs at the bluff body corners and the vortex excitation and damping by the flame itself 33,34 are neglected. The vortex sheets behind the bluff body are discretized into small vortex elements and the flow field can then be solved by tracing the motions of all the vortex elements using the Lagrangian method.…”

Impact of burner plenum acoustics on the sound emission of a turbulent lean premixed open flame

“…For the time integration, a third-order total-variation diminishing (TVD) Runge-Kutta scheme is used. 31 To capture arbitrary fluid-solid boundaries, a conservative cut-cell technique is used where small cut cells are treated using a flux-redistribution method. 43,44 A combined G-equation progress variable modeling approach by Moureau et al 41 is applied to a threedimensional turbulent lean premixed methane-air flame.…”

“…The G-equation approach is used to describe the motion of the inner-layer temperature contour of the flame, where the contour represents the zero levelset, i.e, the G ¼ G 0 ¼ 0 contour, of a three-dimensional function Gðx; y; z; tÞ. The evolution of this zero level-set contour is defined by the G-equation 31,45…”

“…The G -equation approach is used to describe the motion of the inner-layer temperature contour of the flame, where the contour represents the zero level-set, i.e, the G=G0=0 contour, of a three-dimensional function G(x,y,z,t). The evolution of this zero level-set contour is defined by the G -equation 31,45 where the symbol ·ˇ indicates a variable defined at the filtered flame front location boldxtrue˜true^f such that Gˇtrue(x,ttrue)=G0. The symbol true·˜^ represents variables which are filtered by the surface integral over the resolved flame front, whereas true·˜ represents Favre filtered and true·¯ spatially filtered variables.…”

“…For open flame configurations, however, it is generally assumed that the acoustic spectrum does not exhibit pronounced peaks and the acoustic properties of the inflow boundary condition are often neglected in order to speed up the simulation. [31][32][33] There exist only a few studies for laminar flames in which distinct peaks in the spectral sound pressure distribution are observed. Note that the nature of these peaks can be fundamentally different to peaks observed for turbulent flames since there is no broadband sound emission in laminar flames.…”