Boundary treathment in multi-body sim + radiation

General discussion about Elmer
Post Reply
Saint
Posts: 28
Joined: 19 Apr 2023, 08:15
Antispam: Yes

Boundary treathment in multi-body sim + radiation

Post by Saint »

Hello again,

First of all thank you for your patient, I know I have asked so many basic questions; I hope you can guide me as before:

I am simulating a hot object, inside a box; the bot T is fixed and constant, but the air T and the object T can change.
My goal is to simulate and calculate the T and air flow (natural convection), Especially ! while considering the radiation from the hot surface, and back reflection of other surfaces.

I believe, after some time, other surfaces can also act as secondary radiators, and of course reflection of the walls can increase the T of the hot surface as well.
Image

Usually, engineers will simplify the question and only simulate the air, but before that, I need to make sure about the reaction of boundaries.

Noslip
1:?: normally, when there is a fluid in contact with a solid, the surface will be nonslip. I believe that I only need to apply the noslip condition to the boundary of my fluid, but just to be sure, please let me know if I need to check the nonslip condition at the boundary of my solid as well or not?
(I believe it is not needed, since the navier-stoke is not active for the solid body, but better safe to ask than sorry)

Radiation
2 :?: When I want to have radiation, I need external T to be defined (?)(is it only for idealized or for gray as well?)

3:?: When the outside of my boundary has a constant T, it is fine, but how about when there is a heat flux coming to my boundary?

4:?: How about when two boundaries are starting at the same initial T, but I expect one of the surfaces to get hot fast ?


General Radiation Q
5:?: When two boundaries do not have a fixed T or flux, but they are under effect of other elements in the simulation, how can I apply an emissivity to them, and force Elmer to calculate the reflections and secondary radiation?

6:?: Is the external T, in the boundary condition, like dirichlet T (meaning it will be fixed and overwrite everything through the simulation) or is it like initial T (some starting T and it can be changed after some iterations)
or maybe there is a third option....?

Redundancy Q
7 :?: This is not a short question as it seems, but I am wrried about the redundancies of applying these 3 factors multiple time; I know it is case dependant, but :
7.1 _ nonslip : applying this condition both at the solid boundary and fluid boundary.
7.2 _ emissivity : applying the emissivity value both at the material property and boundary condition after idealized radiation.
7.3 _ temperature : when I have heatflux/constant T from one boundary/surface, entering to another, Do I need to apply it both at my hot object and as an external T in my receiving object ?


Some visualization of my question + some other scenarios are demonstrated with the following pictures:

Image

Image
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Boundary treathment in multi-body sim + radiation

Post by raback »

Hi

You should have just one boundary that is shared by fluid and solid. If the mesh is not continuous over the interface you need to apply some mortar BCs, for example, and you don't want to do that if you have just started with Elmer. So ensuring that mesh is conforming over the BC is the 1st thing.

You need "external temperature" for radiation only if you have idealized radiation model either fully or partly (radiation boundary open). If your system is closed there is not external temperature.

The emissivity is ideally only given as material property to the solid. Never for the fluid. However, sometimes one can have fully insulated boundary without the solid being present. Then you have no other option that to give the emissivity at the BC + define the direction of the normal.

-Peter
Saint
Posts: 28
Joined: 19 Apr 2023, 08:15
Antispam: Yes

Re: Boundary treathment in multi-body sim + radiation

Post by Saint »

1. Thanks for pointing out the mesh issue ! My first goal was to have 3 distinguished parts, so I can give them different materials.
I wonder how I can have 3 parts, but with shared boundaries!??!?!..... I use FreeCAD and Gmsh and my go-to option was the "compound function" in FC... now I have no clue...
:?: maybe the -merge 1e-7 or similar value helps ?????

2. Thanks about the external T. I assumed considering my "air" as the reference, the wall will be an external T.
:?: Am I right to assume that by giving emissivity value to materials, then elmer will naturally consider the radiation inside my system?
the images above

:?: so is it safe to assume that the external T is equivalent of dirichlet T but to introduce the radiation of foreign object, meaning only radiation will be exerted from that, but the convection and conduction will not be introduced to system, and to that end, I will need to set the dirichlet T as well?

3. Thank you about the emissivity. From now on, I will not put the emissivity value in my boundaries when I have solid in proximity.
:!: and only when I have, my fluid and the other surface (which is not modeled) is solid, then I will apply the emissivity as BCs of the fluid surface.
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Boundary treathment in multi-body sim + radiation

Post by raback »

Hi,

Dirichlet condition always dominates other boundary conditions. So if you apply them on the BC you cannot apply any other condition. This is generally true for any PDE. You can only describe the value or flux - not both.

Not the "diffuse gray" radiation model is a slight exception in that it will have an effect even in conjunction with Dirichlet condition for other BCs that are viewing the boundaries. See Eq. (1.17) in
https://www.nic.funet.fi/pub/sci/physic ... Manual.pdf

-Peter
Post Reply