Re: [ff3d-users] Extraction of the unit normal vector, or normal/tangential derivatives.

[Stephane Del Pino]

Dear Mykola,

Le mardi 21 octobre 2008, Mykola Tasinkevych a écrit :
>       Dear Stéphane,
>
>   thank you for your advise!
>
>   As I understand, your penalty term is forcing the  flow field "u" only to
> be parallel to the surface of obstacle, but
>  does not prescribe a specific slip,  "u = *slip* on [ref] ", where *slip 
> *is some  function.
Sorry I did not read carefully your first e-mail.

 >   One, in principle, can  also penalize the variational formula in order to
> impose
>          u = *slip
>   *on the boundary.   Is it also possible to  impose this condition
> rigidly, as a boundary
>   condition?
Yes, It should be possible.  Maybe you will need to "precompute", dx(RHO), 
dy(RHO) and dz(RHO) storing them in a function or a femfunction. I am not 
sure that using them for the boundary condition directly is accepted.

Telle me if it works or not.

Best regards,
Stéphane.