[Nek5000-users] Failure when changing geometry in usrdat2

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

Hi all,

I am currently using Nek to look at the effect of convection in slightly
varying geometries, e.g. I start with a basic grid and then deform it in
usrdat2:
call rescale_x(zm1,-ellip*r_out,ellip*r_out)
param(59) = 1
r_out is the maximum value of the original grid in z-direction
(glmax(zm1,n)) and ellip is some scale factor (usually <1). If I just
use the original grid with no deformation, the whole case runs fine (as
far as i have checked), but as soon as i introduce some deformation
(ellip=0.9999 is enough), the code seems to not converge anymore, e.g. I get
Starting time loop ...

Step      1, t= 1.0000000E-04, DT= 1.0000000E-04, C=  0.000 0.0000E+00
0.0000E+00
              Solving for heat
  Solving for fluid F T T
   Temperature/Passive scalar solution
    1.0000000000000000E-008  p22            1           2
   1     1 Helmholtz TEMP    F:   9.1518E-01   4.0000E+00 1.0000E+00
1.0000E+04
           1    Hmholtz TEMP:      0   9.1518E-01   9.1518E-01 4.0000E+00
           1   1.0000E-04  1.0291E-01 Heat done
    1.0000000000000000E-008  p22            1           1
1     1 Helmholtz VELX    F:          NaN   1.0000E-08 1.0000E-03
1.0000E+04
   1     2 Helmholtz VELX    F:          NaN   1.0000E-08 1.0000E-03
1.0000E+04
   1     3 Helmholtz VELX    F:          NaN   1.0000E-08 1.0000E-03
1.0000E+04
   1     4 Helmholtz VELX    F:          NaN   1.0000E-08 1.0000E-03
1.0000E+04
(...)
  1   100 **ERROR**: Failed in HMHOLTZ: VELZ          NaN NaN   1.0000E-08
     1 1.00000E-06         NaN         NaN         NaN       1 Divergence
     2 1.00000E-06         NaN         NaN         NaN       1 Divergence
     3 1.00000E-06         NaN         NaN         NaN       1 Divergence
     4 1.00000E-06         NaN         NaN         NaN       1 Divergence
(...)
DT/DTCFL/DTFS/DTINIT   0.100E-01   0.500E-01   0.000E+00   0.100E-01
Step      2, t= 2.2000000E-04, DT= 1.2000000E-04, C=******* 7.9370E+00
7.9370E+00
              Solving for heat
  Solving for fluid F T T
    1.0000000000000000E-008  p22            2           2
           2   2.2000E-04  4.9921E-02 Heat done
    1.0000000000000000E-008  p22            2           1
    1.0000000000000000E-008  p22            2           1
    1.0000000000000000E-008  p22            2           1
          2 U-PRES gmres:    120         NaN  1.0000E-06         NaN
1.6777E+00  4.4202E+00
            2  DNORM, DIVEX NaN                       NaN
           2   2.2000E-04  4.5553E+00 Fluid done
  numrec=          10
  WARNING: DT<0 or DTFS<0
           Reset DT
and so on for the following time steps. The max/min velocity seems to
be  +- 0.990000E+22.
I have run similar cases without a temperature field before (mechanical
forcings) without any problems. Is there anything special I have to
change when adding a temp. field?

Cheers,
Jan