In a calculation (a jet into river), it was found that a good solution is avaialble at large turbulent prandtl number (about 2). Is this reasonable?

Thanks

Ger

To find a good solution, one cannot, rather should not attempt hit-and-try, playing with turbulent Pr.There is a consideration of 'physics' for the apparently unknown phenomenon. To start, you may go thru the relevant portion of text in the book by Kays and Crawford (Mass and momentum Transfer ?). There was an enlightening discussion in a paper by Kays (?) titled 'Turbulent Prandtl Number - Where are we?'.There have been some papers on estimation of Prtur thru DNS also.

PKM,

Your point is interesting. But, from your knowledge, either from your own experience or literature, how large the number can be?

When Prandtl number > 1, then temperature has a smaller diffusion coefficeint in comparison with diffusion coefficent for velocity fiield. THis means temperture diffuses slower than velocity filed. But, this should not be true in many cases. Right?

Thanks

Ger

Turbulent Prandtl number can be much larger than 1.0 in near-wall region. I recall that one of the Kay and Crawford's papers have a plot for turbulent Prandtl number as a function of y+. Turbulent Prandtl number can be computed from Reynolds stresses, Reynolds flux, and mean velocity gradients.

Ger,

your second para talks about Pr - not Pr-tur. For Pr>1 fluids, during parallel flow, it is the thickness of the boundary layers - (momentum > thermal) which governs the heat transfer mechanism due to normal heat diffusion and eddy diffusion (for turbulent regime).However, Pr-tur being an outcome of the flow situation (that's why y+ comes into picture), needs to be estimated. This is difficult and probably no conclusive, or even indicative reference is so far available. Thus Pr-tur should ideally be taken as function of y+. For evolving flow field during numerical computation this often leads to 'numerical hazard'. It is perhaps better to take Pr-tur as ~1.0 at start and once the approximate flow field sets in, variation of Pr-tur wrt y+ may be incorporated. Going back to your stated problem (impinging jet), one has to be careful about the stagnation effect and appropriate physics shoud be brought in - instead of playing with Pr-tur.