Hello,

I am interested to know if when the G-Equation (Re-initialisation) is applied to a fully- premixed combustion problem in LES, can this model deal with merging of flamelets and the formation of closed pockets of unburnt gas being propogated into a burnt gas region and its burning there. I assume the G-Equation cannot take account of this since it is a distance funtion and therefore will not give instantaneously accurate results. Does anyone know if this limitation is likely to matter in the mean?

Any other comments would be appreciated.

Thanks.

Does anyone have any comments? I would have thought the combustion people might have something to say? I realise that the flamelet assumption (thin reaction sheet) is used for the G-Equation, but even in moderately wrinkled flames, pockets are formed and I don't think the G-Equation approach can deal with this.

As far as I understand the flamelet assumption is not a limitation for pockets. If the pockets are bounded by very thin flame fronts you can base your theory on the flamelet assumption. The flamelet assumption is not only the basis for the G-equation model but also for the TFC of Zimont and other models.

A nice phd-thesis concerning with G-Equation and LES you will find at: http://www.itm.rwth-aachen.de/Deutsc...icht_main.html

Regards MCPDF

Thank you MonteCarloPDF. I have looked at Marcus Hermann's thesis, but this does not deal with LES applied to the G-Equation. Could you name which thesis you are refering to? Thanks for your help, Hermann's thesis is of some interest also.

I agree that the flamelet assumption does not exclude the existance or modelling of pockets. However, I think the G-Equation approach in effect does. Is this correct? If not, why? Thank you.

As my nick says, I have less experience with G-Equation. May be somebody else can help? I would be interested in the answer, also.

Yes, It would be interesting if someone else could elaborate for us! It seems that in Hermann's thesis he has instantaeous plots of temperature which suggest he is capturing 'pockets' of unburnt gas floating into the burnt gas. I would like to know how reinitilsation works for these problems if that is the case.

Also, MonteCarloPDF ... can you explain to me what MonteoCarlo PDF methods are. How do they work, how do you apply them to fluid flow problems. If you can't be bothered to answer, understood.... I reckon it must be a difficult subject to explain!

It is possible to derive a transport equation for the joint probabilty density function of velocity and composition. You can solve this equation via stochastic methods. The great advantage is that the chemical source term in the PDF transport equation is closed!!!! The key problem of combustion simulation seems to be solved. Unfortunately you are hit by some other problems if you use this method. First it is easy to realize that mass is transported in physical space, but have you ever transported probabilty in physical and probability space? Well, such ideas may be funny for some mathematians living in world of formulars. However they are not so amusing for the common engineer.

Beyond this in the PDF transport equation a new unclosed term is appearing. Mixing has to be modeled. This is comparitively simple for diffusion flames, but a difficulty for premixed flames. There are some other issues with the MonteCarlo-PDF method. In my opinion these issues are mostly a question of man and computer power.

Hi Interested, I suppose this discussion would fit better into some forum for combustion science. Do you know such a forum? Regards MCPDF

Hi MCPDF,

No, I am afraid I don't know of such a forum. I don't know if their would be sufficient demand for one. What do people think? Where woudl it be hosted? I don't think CFD-ONLINE.COM would be interested in such a venture as their are many other topics of much more interest to the general cfd-online reader.

Interested.