[Nek5000-users] gmers solver issue

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

Hello,

I trying to run the DNS of a pipe-like configuration with NEK-5000  but I'm facing some issues that I cannot fully understand.

More specifically, I am using the PN-PN formulation and the simulation slows down drastically after certain time. I have observed that when
this  happens the number of GMRES iteration for the pressure solver increases (from 50 at the beginning of the simulation to 170)
and the norm of the divergence of the velocity increases and I get a WARNING DIV(V)-QTL to large (~ 1.2E-1).

The problem appeared initially using a 3rd order polynomial, but even using higher order (9 th) does not solve the issue.

Do you have some hints about the nature of this issue?


Kind regards,
Dante

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20180516/282f9e40/attachment.html>