Baldwin-Lomax
 

Hi Everybody,

I'm working with the Baldwin-Lomax turbulent model in wall bounded flow.

I have a problem in the evaluation of the of the fmax quantity (maximum of the distance from wall * vorticity * damping function). It seems that there is multiple local maximum.

I know that modifications to the Baldwin-Lomax model exist to prevent this problem but i don't know wich modification can interset me and where i can find reference to it.

I try a little googling 'Baldwin-Lomax + modification'. Degani-Schiff modification appears several times but i couldn't find any explanation nor complete reference. Is it this modification that i need ? What is the reference to this paper ?

Thanks in advance for your response or comment

jf

Re: Baldwin-Lomax
 

Just take the largest (global maximum) value

Re: Baldwin-Lomax
 

I have found this by Google at http://people.nas.nasa.gov/~smurman/AIAA2001_0114.pdf and guess the following is the reference you are looking for:

[2] Degani, D., and Schiff, L. B., "Computation of Turbulent Supersonic Flows About Pointed Bodies Having Crossflow Separation," Journal of Computational Physics,vol. 66, pp. 183–196, Sept. 1986.

You may find also references in Degani's homepage http://meeng.technion.ac.il/Personne...gani/data.html.

I hope this helps.

Re: Baldwin-Lomax
 

Yes, it's the reference I sought. It seems my google was tired ;-).

Thank you

Re: Baldwin-Lomax
 

Thanks for your response Tom. However, in my case taking the global maximum of f = w * y * damping (w is the vorticity and y the distance from wall) gives strange results. This case is the 2D flowpath of a compressor delimited by the hub and the shroud. A boundary layer at the inlet of the domain exists and is modeled by a distribution of total pressure.

With this inlet boudary layer, there are 2 peaks of the quantity f. At a certain location, the first peak becomes greater than the second, the value of the maximum of f (fmax) doesn't vary much but the value of ymax (location of fmax) does, say by a factor 4. So finaly, the eddy viscosity wich is proportionnal to ymax*fmax is dividide by 4 for two consecutive cells. It's a least strange.

I think the implementation is clean (validated on a flat plate) and this problem is only related to the (quite large) inlet boundary layer (there is no problem when there is no boundary layer at inlet).

Have you any idea how i can overcome this problem. I must admit that, except the book from Wilcox, i have not read a lot about Baldwin-Lomax. Could you point me reference on pratical utilisation of this model ? Also, have you an idea of how to deal with ''mixing'' boundary layer (when the edge of a boundary layer of a wall ''touch'' the edge of the boundary layer from the opposite wall) ?

I hope I've been clear enough and thanks in advance for your comments.

jf

Re: Baldwin-Lomax
 

It's a long time (>10 years) since I've looked at Baldwin-Lomax and then the only reference I used was the original paper - although I implemented Baldwin-Lomax I used the Cebeci-Smith model in nearly all my calculations.

One option is you could take the maximum nearest the wall - although the solution may look sensitive to how you choose this value in practice the error in the closure compared with reality is much greater. (Personally I'd use a different closure - one that ensures that the stresses are observer independent).

hope this helps (but it probably doesn't)

Tom.

P.S. If I recall correctly the standard Baldwin-Lomax scheme has a problem where the flow separates and so the model needs to be changed just to ensure that the numerics converge.

Re: Baldwin-Lomax
 

> Personally I'd use a different closure - one that ensures that the stresses are observer independent

Do you know such a model ? Could you give some references ?

Again, do you have an idea of what to do when boudary layers from two opposite wall grow and then meet each other ? Someone has an idea ?

Thanks

jf

Re: Baldwin-Lomax
 

From what I can tell from your problem all that is happening when the "boundary layers" merge is that your flow is developing into something akin to the fully developed pipe flow solution. I suspect your problem with the Baldwin-Lomax model is arising because of this - it was designed for external aerodynamics and not internal flows.

You should probably try using a k-epsilon model for this flow instead of Baldwin-Lomax. Alternatively you could make up your own mixing-length formula,

Tom.

Re: Baldwin-Lomax
 

I will think about that.

Thank you again for your inputs.

JF