[Nek5000-users] conj. ht. transfer

[nek5000-users at lists.mcs.anl.gov]

Andrea,

Let's start with an important note: The implement SEM is locally not
conservative. Thus, the fluxes do not need to match. However, it would
be suspicious if the flux conservation at the interface is
proportional to something!

I'll have a look at it.
Stefan

On 10/16/11, S K <stgeke at googlemail.com> wrote:
> Hmm, that puts some pressure on me. Investigation starts now :)
> Cheers,
> Stefan
>
> On 10/16/11, nek5000-users at lists.mcs.anl.gov
> <nek5000-users at lists.mcs.anl.gov> wrote:
>>
>> Dear Andrea,
>>
>> Thank you for your careful investigation of this issue.
>> We'll dig into it in the next 36 hours and hopefully we
>> will get the problem resolved quickly.
>>
>> Best regards,
>>
>> Paul
>>
>>
>>
>>
>> On Sun, 16 Oct 2011, nek5000-users at lists.mcs.anl.gov wrote:
>>
>>> Dear all,
>>>
>>> I was running a simple case with conjugate heat transfer. I managed to
>>> validate it against some analytic data.  I checked whether the heat
>>> fluxes
>>> at
>>> the interface between solid and fluid match or not. For any ratio of
>>> solid
>>> conductivity over fluid conductivity they match very well especially if
>>> the
>>> grid is very refined at the interface. The problem is that they match
>>> only
>>> if
>>> vtrans of solid is the same as vtrans of the fluid, else the ratio of
>>> the
>>> two
>>> fluxes is proportional to the ratio of vtrans, which is wrong.  I had a
>>> look
>>> into the code even though I am not very expert. I am using cvode. In
>>> fcvfun
>>> there is a call to makeq. In makeq, after the call to wlaplacian and
>>> add2,
>>> the rhs, which is bq, is divided by vtrans. When we are back to fcvfun,
>>> there
>>> is a call to dssum over ydot. I am not sure but I think that dssum mixes
>>> the
>>> properties of fluid and solid, therefore the transient equation which is
>>> solved for the temperature at the interface has a flux which is divided
>>> by
>>> vtrans of the fluid and another flux which is divided by vtrans of the
>>> solid,
>>> therefore the two fluxes can not match at the interface unless solid and
>>> fluid have the same vtrans.  If this is really the problem, I think it
>>> could
>>> be solved by dividing bq at the interface by vtrans of the solid only,
>>> before
>>> dssum is performed.  I'll check again my case to be sure that what I
>>> observed
>>> was not due to an error.
>>>
>>> Thanks in advance for your help.
>>>
>>> Regards, Andrea.  _______________________________________________
>>> Nek5000-users mailing list Nek5000-users at lists.mcs.anl.gov
>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>>>
>> _______________________________________________
>> Nek5000-users mailing list
>> Nek5000-users at lists.mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>>
>