[aspect-devel] Velocity Boundary conditions

[Ian Rose]

Hi Lev,

I was probably over-zealous in my comment. After all, zero-velocity
boundary conditions are specifying all three components of the velocity. I
just meant that if you specify all three, you need to be very careful not
to violate incompressibility (imagine a pure divergence velocity field
where all arrows point away from the origin). Otherwise there will be no
valid solution and the solver will run into issues.

Ian

On Sat, Jun 3, 2017 at 10:25 PM, John Naliboff <jbnaliboff at ucdavis.edu>
wrote:

> Hi Felipe,
>
> This indeed a bit of an odd setup and it would probably be helpful to see
> the exact benchmark you are trying to reproduce.
>
> To have vertical flow in and out of the bottom boundary you need a
> traction boundary condition rather a velocity boundary condition. However,
> I’m not sure if you can mix traction and velocity b.c.’s on a single
> boundary?
>
> I’m pretty sure the answer is no and I can’t recall a study in the
> literature that did this. There may be a way to implement this in ASPECT,
> but as you noted this is definitely not a “standard” setup.
>
> Cheers,
> John
>
> *************************************************
> John Naliboff
> Assistant Project Scientist, CIG
> Earth & Planetary Sciences Dept., UC Davis
>
>
>
>
>
>
> On Jun 3, 2017, at 6:18 PM, FELIPE ORELLANA ROVIROSA <
> f_orellana at berkeley.edu> wrote:
>
>
>  Hi all,
>
>    uf, now back in the saddle.
>
>     On the lower boundary (horizontal plane surface with normal z) of a 3D
> system, I am trying to have velocities that in a Cartesian form are like
> this:
>
>     V = ( 0 , 0 , Vz )  =  0 (x) + 0 (y) + Vz (z)
>
>     with Vz determined by the Stokes, Continuity and Energy equations, and
> the Temperature at that boundary. This also means that I cannot
> write/specify a value for Vz there.
>
>     Interpreting this, I want to preclude horizontal velocities on that
> surface, but I want to have a free flow (in/out) across it.
>
>     The difficulty is how to write this in the Aspect parameter file, if
> it is possible in the context of how the equations are solved internally.
>
>     Aspect seem to impose zero net mass-flow on a boundary whose
> velocities are unspecified, I have tested this, and it works OK. So, now, I
> would also expect to have zero net flow across the boundary.
>
>   I know that this type of setting might be a perhaps unphysical, but I
> want to try it regarding a benchmark I am doing, against a published (1994)
> model.
>
> Thanks to you all,
> cheers,
> Felipe
>
> On Sat, Jun 3, 2017 at 1:31 PM, John Naliboff <jbnaliboff at ucdavis.edu>
> wrote:
>
>> Hi Felipe,
>>
>> In your function you would need to set values for all three components,
>> but in the "Model settings subsection" you specify what components of the
>> velocity are applied to each boundary under "Prescribed velocity boundary
>> indicators".
>>
>> For example, if you want the function to only apply to the "x" velocity
>> component you would write:
>>   set Prescribed velocity boundary indicators = left x: function, ....
>>
>> The "y" and "z" components of the velocity field are then not fixed along
>> this boundary (i.e. parallel to the boundary is free-slip).
>>
>> A full description of this is located in the "Prescribed velocity
>> boundary indicators" section in the output parameter file or in
>> source/simulator/parameters.cc.
>>
>> Cheers,
>> John
>>
>> *************************************************
>> John Naliboff
>> Assistant Project Scientist, CIG
>> Earth & Planetary Sciences Dept., UC Davis
>>
>> On 06/02/2017 03:04 PM, FELIPE ORELLANA ROVIROSA wrote:
>>
>>
>>  Hi Wolfgang,
>>
>>      Wow, I am amazed. So, that means that I should prescribe two
>> components and leave the third in omission.
>>
>> for a wall with normal vector z
>>
>>       Vx =0 , Vy=0
>>
>>
>> how to write this in the parameter file?
>>
>> Set them as prescribed velocities? and then prescribe just two components?
>>
>> subsection Boundary velocity model
>>   subsection Function
>>     set Variable names      = x,y
>>     set Function expression = 0 ; 0 ;
>>  end
>> end
>>
>> ??
>>
>> Felipe
>>
>> On Fri, Jun 2, 2017 at 2:26 PM, Wolfgang Bangerth <bangerth at tamu.edu>
>> wrote:
>>
>>> On 06/02/2017 03:17 PM, FELIPE ORELLANA ROVIROSA wrote:
>>>
>>>>
>>>>         At a wall boundary, is it possible to prescribe 2 components of
>>>> the velocity (for example to 0 in particular for my case), and leave the
>>>> 3rd free?
>>>>
>>>
>>> Only if those two are the two tangential components of the velocity.
>>>
>>> Best
>>>  W.
>>>
>>> --
>>> ------------------------------------------------------------------------
>>> Wolfgang Bangerth          email:                 bangerth at colostate.edu
>>>                            www: http://www.math.colostate.edu/~bangerth/
>>> _______________________________________________
>>> Aspect-devel mailing list
>>> Aspect-devel at geodynamics.org
>>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>>
>>
>>
>>
>> _______________________________________________
>> Aspect-devel mailing listAspect-devel at geodynamics.orghttp://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>>
>>
>>
>> _______________________________________________
>> Aspect-devel mailing list
>> Aspect-devel at geodynamics.org
>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>>
>
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>
>
>
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20170606/19f026bd/attachment.html>