Re: [sparta-users] [EXTERNAL] 2d axi with global weight cell radius/only + adapt_grid

[Gallis, M. A. <ma...@sa...>]

Pierre

Grid adaptation should work in 2D-axisymmetric. However, you must be careful when you adapt not to be left with cells with very few particles or with timesteps that allow particles to cross multiple cells in one timestep. Both of these will increase the discretization error despite decreasing your cell size.

One easy way to verify that this is your problem is to increase the number of simulators and decrease your timestep. If this makes the problem go away it is the way you are adapting,

On what criteria are you adapting the grid?

Michael

From: pi...@rt... <pi...@rt...>
Date: Friday, September 13, 2024 at 9:49 AM
To: spa...@li... <spa...@li...>
Subject: [EXTERNAL] [sparta-users] 2d axi with global weight cell radius/only + adapt_grid
Some people who received this message don't often get email from pi...@rt.... Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Dear SPARTA users,

Summary:

I am testing the axisymmetric capabilities of sparta on a 1m diameter sphere re-entering from LEO, around 110km of altitude. While I was able to carry some successful cross-code validation in 2D and 3D, I am struggling with the axisymmetric mode when combined with the grid adaptation of sparta.

Description of the problem:

The results at the wall over the 1m diameter sphere re-entering from LEO are plotted in the attached file:
- Cf: friction coefficient
- Cp: pressure coefficient
- Cq: heat coefficient
In my setup I am using: "global weight cell radius/only". I made sure my results converged regarding the time step and the number of particles.


1) If I do not make any calls to "adapt_grid" in my script, the results are OK-ish: Cp_max ~ 2 at the stagnation point. But this is only because my initial grid is adapted manually: 1 mean free path everywhere. and 0.25 mean free path at 0 distance from the sphere

2) Obviously, I do want to use "adapt_grid" to get a better match with the ref. However, when I do use "adapt_grid" 2 or more times, I obtain a maximum stagnation pressure coefficient of ~8. That is ~ 4-5 times higher than I am expecting. More specifically the pressure in the area where the cells have been adapted is much higher than expected. I also get unrealistically high heat rates coefficients (Cq) at the stagnation point: it is 3.5 times larger than 1 : so higher than in free molecular flow. In attachment is a picture of the grid at the end of the simulation for illustration.

My guess is that the adaptation (adapt_grid) is not compatible with "global weight cell radius/only".

The question:

What strategy shall I use to carry out an axisymmetric computation combined with grid adaption ?

Thank you in advance and best regards,

Pierre