Keywords: direct numerical simulation, parallel computing, turbulence, lifted jet flame, combustion, flame surface.

DNS of a real turbulence combustion system is still very rare and a big challenge nowadays. Hence, we plan to perform a DNS of a real combustion experiment and expected to compare the DNS results with the measured data carefully to ensure the fact that DNS can resolve turbulent combustion indeed since 2009. With this aim in mind, an experiment carried out by Cabra [1] was chosen to be modelled using the DNS method. A strict grid of points fulfilling the Kolmogorov scale can have up to 6 billion points for the Cabra experiment, so a stretch grid system is employed to reduce this number. We have introduced our stratagem of stretch grid designing method in the paper. With this grid system, the number of grid points reduces to 285 million but still preserves the ability of resolving the Kolmogorov scale in regions of concern.

The efficiency of the large-scale computation is only 49% compared with the test case but it is still within our tolerance. It has been computed upto 12.2 ms in real time which is long enough to obtain a statistical data. The comparison between our DNS results and measured data is excellent not only in Favre averaged scalars but also in corresponding r.m.s. scalars. It can then be further concluded that DNS can totally resolve turbulence combustion if proper inflow conditions and reaction mechanics are involved.

An instantaneous temperature field is also shown in the paper. It can be seen that the temperature field is irregular, which is a result from the interaction between the turbulence and the flame surface.

1

R. Cabra, "Turbulent Jet Flames into a Vitiated Coflow", PhD. University of California, Berkeley, 2003.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £85 +P&P)