[FLASH-USERS] How to describe vacuum in cfd simulation?

[Ryan Farber]

Hi Sang Yun,

As far as I'm aware, the typical way to represent a vacuum in fluid
simulations is with a very small density (effective vacuum) since absolute
zero density is problematic regarding stability (as you've noticed).

Perhaps another method would involve being creative with a user defined
boundaries. That is, setting BDRY_VAR = 1 in regions where the vacuum is
supposed to be. It might be okay to set density to zero there if the fluid
solver skips boundary values. This might be okay if you're simulating water
in a lake and you want the air to be a vacuum and you're ignoring
evaporation of water. However, in most cases what I've described isn't
suitable. For example, if you're modeling a supernova exploding into the
interstellar medium, then you'd want the initially "vacuum" values to
receive gas from the supernova. However, in this case, there isn't a true
vacuum to begin with since the interstellar medium does have particles in
it.

Best,
--------
Ryan


On Thu, May 9, 2019 at 11:53 PM sang-yun Shin <shinsy85 at gmail.com> wrote:

> Dear FLASH users,
>
> I want to describe some fluid as vacuum in cfd simulation.
> When I described vacuum in particle-in-cell simulation, it was easy with
> concept "there are not any one."
> So with this same concept, I made a fluid with zero density.
> But, after such setup, simulation does not work. (zero energy ? error
> occurs.)
> Fortunately, with replacing a zero density to few density (~10^-6 g/cm^3),
> simulation work well.
>
> MAIN QUESTION)
> Any other nice way to describe vacuum in cfd simulation?
>
>
> Best regards,
> Sang Yun Shin.
>
>
>
> ## 신 상윤 (Sang Yun, Shin)
> ## 06974 서울특별시 동작구 흑석로 84
> ## 중앙대학교 104관 (자연과학대학, 수림과학관) 7층 704호 (이론물리연구실)
> ## Tel) None
> ## Mobile) 010-3687-9486
> ## E-mail) shinsy85 at gmail.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20190510/435d31b7/attachment.htm>