Modelling extinction and reignition in turbulent flames

The presented work attempts to extend the conditional moment closure method for noon-premixed. turbulent combustion to predict extinction and reignition phenomena in turbulent flames. The conditional moment closure method is one of a class of conserved scalar modelling approaches in turbulent non-premixed combustion. where chemistry is treated as mainly dependend on the mixing of oxidizer and fuel. However. as designers of combustion devices aim for higher turbulence rates to enhance mixing and promote combustion, chemical conversion is not solely determined by the rate at which fuel and oxidizer are mixed, but kinetic effects become important. Therefore it is necessary in these cases. to consider a second variable to govern the evolution of the chemical system. This variable will parameterize the chemical conversion process from cold. mixed reactants at fixed eguivalence ratio to an eguilibrium state. Equations describing the chemical system as a function of these two variables, the conserved scalar, commonly referred to as mixture fraction and the progress variable. can be derived and constitute the doubly conditioned moment closure equations. However, solution of this set of equations is computationally expensive and key parameters describing the rate of dissipation of the progress variable, which is a reactive scalar, are not yet fully understood. By considering conditional fluctuations of the progress variable, applying simple relationships for scalar dissipation and using a pre-computed functional dependence of conditional moments on the progress variable, the effect of double conditioning on the chemical source term and on the overall chemistry predictions can be examined. The methodology is tested for its capability to predict the turbulent. piloted flames of the Sandia D-F series. These laboratory flames show an increasing degree of local extinction and reignition due to varying turbulence levels. Hence they provide an ideal benchmark for the study of models trying to predict these phenomena.EThOS - Electronic Theses Online ServiceGBUnited Kingdo