open TELEMAC-MASCARET The mathematically superior suite of solvers

I would like to specify a no-slip condition at the solid wall boundaries in my model(friction at the walls). I know that the AUBOR variable found in the conlim file represents the friction coefficient at the boundary if LIUBOR and LIVBOR = 2 and that the default is 0 (slip condition). However, based on the info in the user manual, it isn't clear to me what value I should assign to AUBOR in order to specify the desired roughness at the walls.

I am using LAW OF BOTTOM FRICTION = 4 (Manning's) and am not using the k-Epsilon model.

Thanks!

Hello

While you are not using the k-eps model, you need to introduce the value of AUBOR to consider friction at the solid boundary. This is independent of the friction low you are using for the bottom. The influence of such parameter is computed like dU/dn=AUBOR*U (similar for V).
Physically, to consider a friction effect, AUBOR has to be negative.

I hope that this helps

Riadh ATA

Thanks for the response but I am still unsure what negative value I should assign to AUBOR. Could you give me an example? What value would I assign to AUBOR in my .clm file if I wanted to specify a wall roughness of n = 0.03?

Hello,

You just have to add in your steering file :

LAW OF FRICTION ON LATERAL BOUNDARIES : 4 (this is for Manning)
ROUGHNESS COEFFICIENT OF BOUNDARIES : 0.03 (the constant Manning you want)

Regards,

JMH

Thank you Jean Michel. But wouldn't that apply the no-slip condition to all the lateral boundaries? Do I not need to distinguish between the slip (external boundaries) and no-slip (internal) boundaries in my model by changing the AUBOR value in the .cli file? If so, what value do I assign to AUBOR?

Thanks for your help!

Hello,

That will apply to all the solid boundaries, but not the liquid boundaries. Otherwise it is difficult to find directly AUBOR, because :

NUT * DU/DN = AUBOR*U + BUBOR, where NUT is the turbulent viscosity, as N is the exterior normal we see that AUBOR must be negative to get a friction, but it is not straighforward to link it to a friction coefficient without using a logarithmic wall theory with a rough turbulence regime. If you want friction on some solid boundaries and not on others, it would be a matter of cancelling AUBOR for them, and this would be a small development (do it in bord.f for example, provided it is not rebuilt after).

With best regards,

Jean-Michel Hervouet