[CIG-LONG] Gale 2.0.0

[Walter Landry]

Rainer Nerlich <rainer at simula.no> wrote:
> Hi Walter,
> 
>  sorry for my late response. I have attached three different input files
> and some synthetic data. The input files vary only with respect to the
> diffusivity and the data they are supposed to call as velocity boundary
> conditions. In other words, the initial temperature conditions as well as
> the temperature boundary conditions are for all three the same and only the
> exponents of the velocity boundary conditions vary. This leads to the
> “strange” time step results I wrote about in my last email.
> 
>  I also noticed that if I keep the velocity conditions fixed but vary the
> temperature conditions, the time steps won't change at all. Since you told
> me, though, that the time steps are also calculated based on the
> temperature conditions, that shouldn't be the case...

If I run the three input files, the time step scales as expected.

  $ ../build/bin/Gale 10_23_velocity_temp_BCs_temp_ICs.json 
  TimeStep = 0, Time = 0
  TimeStep = 1, Time = 7.25777e-19

  $ ../build/bin/Gale 10_22_velocity_temp_BCs_temp_ICs.json 
  TimeStep = 0, Time = 0
  TimeStep = 1, Time = 7.25777e-18

  $ ../build/bin/Gale 10_21_velocity_temp_BCs_temp_ICs.json 
  TimeStep = 0, Time = 0
  TimeStep = 1, Time = 7.25777e-17

Is this what you see?  However, you did not scale the viscosities in
the input files, so I would expect different results when the
viscosities become relatively small.

Cheers,
Walter Landry