Dear Niki,
similar to Will I'm trying to use PyFR to simulate incompressible,
turbulent pipe flow. As for the nature of pipe geomtries the simulations
would greatly profit from a streamwise cyclic boundary condition. Hence my
question: Would your workaround of enforcing a constant pressure source
term in the negdivconf.mako file work for a 3d case as well? Would a
pressure drop in z-direction be achvied by adding for example "tdivtconf[3]
= tdivtconf[3] - 4.0;"?
Do you plan to implement such a streamwise cyclic boundary condition in
PyFR in general?
Thanks in advance and kind regards,
Hendrik
Am Dienstag, 21. November 2017 18:57:24 UTC+1 schrieb Niki Loppi:
>
> Hi Will,
>
> If you look at the pressure field early in the simulation you can clearly
> see that there is a periodicity issue. The source term you impose is
> perfectly fine, but it does work for the gradient between the periodic back
> face and the front face.
>
> p(x) = 0 - 4x,
>
> p'(x)= -4
>
> but derivative at the discontinuity reads p'(0) = [p(0) - p(8)]/dx = [0 -
> (-32)]/dx.
> Currently, we do not have a functionality to force the constant pressure
> gradient in a periodic domain. However, for this case you can add the
> following line to pyfr/solvers/baseadvec/kernels/negdivconf.mako, which
> adds a "pressure derivative" source term (-4.0) to the x-momentum equation
> (index [1]).
>
> % for i, ex in enumerate(srcex):
> tdivtconf[${i}] = -rcpdjac*tdivtconf[${i}] + ${ex};
> % endfor
> + tdivtconf[1] = tdivtconf[1] - 4.0;
> </%pyfr:kernel>
>
> I did not properly test it, but I have attached a picture what I got after
> running for a minute. Please note that the pressure looks constant if you
> add the source term this way, because p = du/dx + dv/dx = 0 + 0. I also
> suggest using BDF2, rk4 or tvd-rk4 and multigrid.
>
> Regards,
> Niki
>
>
>
>
>
>
> On 21/11/17 11:30, Will wrote:
>
> Dear Niki,
>
> Thanks for your advice. I have got good alignment on cavity flow now.
>
> I am now trying to simulate pressure driven laminar 2d pipe flow with
> periodic inlet/outlet BC. The velocity profile at roughly 50 seconds is
> still not parabolic. Instead, the flow generates some vortex at ReD less
> than 200 when ac-zeta = 20. The result is attached also the .pyfrm and .ini.
>
> The pressure gradient is constant 4 and nu = 0.01.
>
> Best regards,
> Will
> --
> You received this message because you are subscribed to the Google Groups
> "PyFR Mailing List" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected] <javascript:>.
> To post to this group, send email to [email protected]
> <javascript:>.
> Visit this group at https://groups.google.com/group/pyfrmailinglist.
> For more options, visit https://groups.google.com/d/optout.
>
>
>
--
You received this message because you are subscribed to the Google Groups "PyFR
Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send an email to [email protected].
Visit this group at https://groups.google.com/group/pyfrmailinglist.
For more options, visit https://groups.google.com/d/optout.
