[Nek5000-users] Apply boundary conditions form a restart file

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

Dear Nek users,


I am modeling a laminar boundary layer over a flat plate which will be used as inlet condition for a larger problem afterwards. I got accurate results by letting the boundary layer develop from a uniform velocity boundary condition. Right now I want to truncate the domain, such that the boundary layer is modeled from a point downstream of the flat plate's leading edge.


So far I've managed to interpolate the solution from the complete domain to the truncated domain using grid-to-grid interpolation (subroutine g2gi). I saved the interpolated solution a file called 'new0.f00001'. When I start my new simulation (on the truncated domain), I load 'new0.f00001' on the first timestep (istep = 0) using load_fld() in userchk(). Finally, I copy the initial field to a common variables such that I can access them at every timestep in userbc(), where I impose the appropriate values.


I outpost the initial field (in istep=0) and everything looks fine when plotted on VisIT or examined using hpts(). I've also examined 'ux' and 'uv' from userbc() by writing the values to an external file and again everything looks correct. However, when it's time to solve the Helmholtz system I get this:


"

Starting time loop ...

     DT/DTCFL/DTFS/DTINIT   0.450E-01   0.000E+00   0.694-309   0.450E-01
Step      1, t= 1.7500450E+03, DT= 4.5000000E-02, C=  0.242 0.0000E+00 0.0000E+00
 Solving for fluid F T T
   1.0000000000000000E-008  p22            1           1
  1     1 Helmholtz VELX    F:          NaN   1.0000E-08   2.5238E-04   2.2222E+01
  1     2 Helmholtz VELX    F:          NaN   1.0000E-08   2.5238E-04   2.2222E+01
  1     3 Helmholtz VELX    F:          NaN   1.0000E-08   2.5238E-04   2.2222E+01
...
"


Could anyone suggest a reason why this is happening? The field values seem to be loading just fine, and I think the boundary conditions are being applied correctly, so I don't understand why I would get NaN.


Thank you.


Juan Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160323/3a905ea2/attachment.html>