An efficient PDF calculation of flame temperature and major species in turbulent non-premixed flames

In spite of recent developments in the PDF calculations of turbulent flames, the high computational time required to implement PDF simulations makes it intractable in practical applications. Therefore, it is important to design and select different parameters for PDF calculation of most important quantities, i.e. temperature and major species means, in an efficient manner. The ingredients of the present model are a standard k-epsilon turbulence closure for modeling flow field and a joint composition PDF closure for the scalar fields. A modified Curl model is applied to consider molecular mixing in PDF transport equation and a simplified two-step mechanism which lowers the computational cost is incorporated to describe the chemistry. The flow field is solved numerically using an upwind discretization for the convective terms and a central discretization for the diffusion terms by coupling it with an Eulerian Monte Carlo algorithm to solve PDF transport equation. To show the superiority of the current PDF calculations over traditional moment-closure methods commonly used in practical applications, simulation is also performed by RANS method which shows large discrepancies, especially in prediction of maximum flame temperature (on the basis of present results, predicted flame temperature has 26% error via RANS method and 8% error via PDF method). Stoichiometric flame length predicted by RANS has 10% error while, by PDF method, this error is negligible and about 0.6%. The effect of coefficient C-phi on the modified Curl model is also investigated and it is concluded that the commonly used value C-phi = 2.0 is the best choice for the case of study. The numerical results obtained reveal that Westbrook-Drier mechanism is working very well in fuel-lean (F < F-st) non-premixed combustion and also it predicts the total heat released in methane combustion in a very good agreement with the experiment. (C) 2009 Elsevier Inc. All rights reserved.