[armyou]

Hi all,

Just want to discuss how to deal with the "solid" in CFX. I know there is an approach to use the high viscosity to freeze the liquid. However, this approach is not very good in my case, because there is still motion in the "solid", although the velocity in "solid" is relatively smaller than in liquid.

Does anybody have other methods?

ps: ghorrocks posted that "some have a moving front which has a sub-domain (or separate domain) behind it", which I totally don't understand. Can some explain it in detail?

Thanks

[ghorrocks]

Don't dismiss the immersed solid approach because it does not sound exact. CFD is a whole pile of approximations and estimates piled on top of each other which, providing the approximations are close enough can simulate reality. Saying that one additional approximation (ie the immersed solid approach) renders the whole thing invalid is rubbish. You just need to make sure the immersed solid approach is valid your case and use it appropriately. Then the error is small enough that the results are useful.

You seem to quote me but I have no idea what it is about. You would have to put in the context of that quote.

[armyou]

Originally Posted by ghorrocks

Don't dismiss the immersed solid approach because it does not sound exact. CFD is a whole pile of approximations and estimates piled on top of each other which, providing the approximations are close enough can simulate reality. Saying that one additional approximation (ie the immersed solid approach) renders the whole thing invalid is rubbish. You just need to make sure the immersed solid approach is valid your case and use it appropriately. Then the error is small enough that the results are useful.

You seem to quote me but I have no idea what it is about. You would have to put in the context of that quote.

Hi ghorrocks,

Thanks for your reply. Here is the address of your reply for another thread: http://www.cfd-online.com/Forums/cfx...sting-cfx.html

Actually my case is about a solid metal melts in a liquid pool. I don't know what is "immerse solid" approach. Would you please explain it further or recommend me some literature about this approach.

Thanks very much.

[ghorrocks]

As I said in the post, search the literature for the current ideas of how to model solidification.

But the basic idea on this approach is to track the solidification front and get the mesh to follow it. Then you have a solid domain behind, a liquid domain in front and some form of interface conditions to handle the solidification. Easy! (Not at all, this is going to be very tricky to implement, and even harder to get accurate.)

[armyou]

Thanks. ghorrocks.

This method seems to be suitable for simulating the rigid solid in liquid. We need to define two domains for the solid and liquid respectively. Is it possible to make the solid to melt in liquid?

[ghorrocks]

Yes, of course. But you will have to develop models for much of the physics yourself as this is not a built-in function for CFX. I think other packages (eg Flow3D) has solidification as a built-in function. I have no idea if it is any good, I have never used it.

[armyou]

Thank you very much, Glenn