[Cartuns11]

hello everybody,

I'm working on a viewFactor model with a complex geometry using openfoam3.0.1. The picture1.png shows it. The geometry consists of 4 parts:

• domain0: atmospheric air
• support: is a solid
• cover: is a baffle
• sourceHeat: solid

I've been carrying out some viewFactor calculations. Using the geometry from picture1, the first results were wrong because the geometry was complex and the viewFactors have some problems on baffles because sometimes these have gaps between the master and the slave and the viewFactors get values 1e+20.

I've found a solution which consists in to close the internal domain using a baffle, the idea is shown in the picture2.png. At the moment, the viewFactors are calculated in two domains and get correct values. Now, my model have 2 fluid regions, but both should be the same fluid....

however, this solution has a handicap, the U, p_rgh and T have to cross the new cover. It is a baffle (master and slave) and I don't know which BC is necessary.
In the first model, the radiation isn't important on the new cover. The picture3.png shows how the BC should work.

I'm using chtMultiRegionSimpleFoam, viewFactors model and v3.0.1.
Could you help me? any idea?

thanks

(sorry for my english)

[avila.vc]

Hi!

As far as I know, to let mass pass through, you could use cyclic boundary conditions. (I am really not sure about that, you could take a look about cyclic boundary, thats all).

However, when I want mass/energy/momentum transport, I dont use boundaries. So, why exactly the first simulation didnt work?

Maybe the second simulation worked because there was no mass transfer between domain 0 and domain 1? That said, maybe the problem was not the need for the new cover... So...could you give some more details about the first simulation? (also checkMesh results, boundary conditions, etc). What solver are you using?

[Cartuns11]

Hi avila.vc,

To sum up, I'm working on a radiation model using the chtMultiRegionSimpleFoam solver. I'm using the viewFactors model to calculate the radiation.

The geometry of the first model, it has been shown the las post, consists of two closed surfaces ( support and sourceHeat) and one stl surface which is the cover, which was defined as baffle. In this case, I had some problems to obtain correct values of viewFactor on the cover due to the mesh, because between the master and the slave, there were some gaps which were very small but also were enough big to calculate the viewFactors........I obtained values like 1e+20. I don't know why, but the snappyHexMesh create gaps on baffles.

One solution was to change the definition of cover, and if the internal volume was closed, it'd be a new cellZone with faceZones. And then, The cover was a faceZone with correct mesh and the viewFactors was correct. This is the reason why I have to close the internal volume with a new cover which is a stl surface.

The newCover is a baffle because the viewFactors aren't been calculated on it.

According to the real model, obviously the internal and external volume are the same fluid, and therfore the new cover allows to pass U, T, p_rgh. At the moment, the radiation isn't important on the newCover.


The problem is the viewFactors on baffles.........and I'm working on soltutions. May be there are others.......

Thanks

[avila.vc]

Hi,

Thanks for the infos

You mean that what you tried to do is to create a cell/face zone at the internal volume? If that is the case, you can do that without closing the volume physically.

I am not used to create meshes with OpenFOAM, I usually create with ICEM(Ansys) or gmsh (free software) and them export to openfoam. Maybe you could also try to create the cover with these softwares that I believe don't have the problem of creating gaps on it.

Best luck

[Cartuns11]

Hi avila.vc,

Thanks for you help. The model has to be carried out with automatic mesh using the snappyHexMesh.

My model have to be two fluid regions and both share a boundary which has to allow the mass flow exchange (U, p_rgh and T)

The boundary is a "mappedWall" type ( it is a faceZone) and I don't know which BC is the correct for U, p_rgh and T.

I'm going to write another thread with a simpler model and another description more detailed.

After, I'll put the link of the other thread with the solution in this post.

Greetings

[avila.vc]

ok.
But I dont think mapped boundary allows mass exchange, it just map the field from a neighbor set of cells.

I really did not understand why you need to divide your domain with a boundary if there is no boundary in the physical problem at that points, specially when you want to have mass, heat and momentum transport through it.

Please let me know when you find a solution.

Best regards