Re: [sparta-users] [EXTERNAL] Re: adapt_grid

[Moore, S. <st...@sa...>]

Yes as Georgii said, you can use "compute lambda/grid" to get mean free path. It may also be good to do some time averaging of the MFP input values, using fix ave/grid, something like this:

compute 1 grid species nrho temp
fix 1 ave/grid 1 1000 1000 c_1[*]
compute 2 lambda/grid f_1[1] f_1[2] Ar

You can also use the dynamic "fix adapt/grid" for adaptation during a run, with a frequency that matches the time averaging from fix ave/grid so you don't have to do multiple static adaptations.

fix 2 adapt 1000 all refine ...

________________________________
From: geo...@dl... <geo...@dl...>
Sent: Sunday, September 18, 2022 3:34 AM
To: Sek...@ng... <Sek...@ng...>; spa...@li... <spa...@li...>
Subject: [EXTERNAL] Re: [sparta-users] adapt_grid


Re: iterate (may be wrong) I think it governs the level of refinement achieved, but if it is set too large and no refinement occurs during an iteration the whole adapt grid loop stops (so it should converge at some point).


MFP adaption is there — a compute call is needed to calculate the MFP in each grid cell, and then a grid adapt based on a compute value can be called. There is an example on the adapt_grid documentation page:

compute 1 lambda/grid c_12 NULL N2 kall
adapt_grid refine coarsen value c_12 0.05 0.1 &
           combine min thresh less more


Mit freundlichen Grüßen
Georgii Oblapenko
Postdoctoral Researcher
Alexander von Humboldt Stiftung Fellow
Institut für Aerodynamik und Strömungstechnik

Deutsches Zentrum für Luft- und Raumfahrt (DLR)

________________________________
From: Ramakrishnan, Sekaripuram [US] (SP) (Contr) <Sek...@ng...>
Sent: 18 September 2022 02:00:13
To: spa...@li...
Subject: [sparta-users] adapt_grid


Is it possible to add an option to adapt_grid based on local mean-free-path?  In my experience with DAC97, that seems to work well.



Rama