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

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

I guess I will answer my own question. It turns out that by loading the file the way I was doing it (calling load_fld() in userchk()), the program tries to calculate the convection terms using previous timesteps, which were not specified in my case. This produces the NaN during the calculation of the RHS.


I solved my problem by setting my *.rea file to:


    1 PRESOLVE/RESTART OPTIONS  *****
new0.f00001

Then in userchk() I copy the initial field to auxiliary common arrays (ub and vb, similar to those used in the "blasius" example) during istep=0 which remain unchanged throughout the simulation and are used to set the appropriate boundary conditions.

Juan Diego

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20160324/134a0550/attachment.html>