Hi!
During silumin casting solidification simulation I found one unpleasant thing. In boundary layer temperature falls lower than initial condition. I understand that it happen due to high temperature gradient. But how to fix it?
Is it mash densifying at boundary layer or smth else? At the 50th step the anomaly desapears.
Here the photos on cloud of:
Initial condition
https://yadi.sk/i/3K9wB_gBnhta9
12th time step
https://yadi.sk/i/PoEADZsjnhtbr
Red arrow points to line of temperature plot.
P.S. After inprovement, I'd like to create a post in "Contributed cases" topic with full description and video.
Casting solidification
-
raback
- Site Admin
- Posts: 4871
- Joined: 22 Aug 2009, 11:57
- Antispam: Yes
- Location: Espoo, Finland
- Contact:
Re: Casting solidification
Hi
Standard FEM formulation just tries to fit the temperature so that it minimizes the L2 norm of the residual. This does not guarantee anything on the monotonicity (there is a very experimental FCT scheme in Elmer that tries to do this). Often it is a sign of too coarse time or space resolution. Also the initial guess may be problematic since it is artificially sharp and does not present any real physical initial state. Sometimes it is possible to go from strong Dirichlet condition to weakly enforce the BCs. Also in this case it might be possible to introduce thermal resistance between different pieces and define heat transfer coefficients between the parts. Unfortunately for complex cases this is probably at least equally problematic. So I would try to improve the mesh around the sharp gradients and think whether the initial state could be made smoother.
-Peter
Standard FEM formulation just tries to fit the temperature so that it minimizes the L2 norm of the residual. This does not guarantee anything on the monotonicity (there is a very experimental FCT scheme in Elmer that tries to do this). Often it is a sign of too coarse time or space resolution. Also the initial guess may be problematic since it is artificially sharp and does not present any real physical initial state. Sometimes it is possible to go from strong Dirichlet condition to weakly enforce the BCs. Also in this case it might be possible to introduce thermal resistance between different pieces and define heat transfer coefficients between the parts. Unfortunately for complex cases this is probably at least equally problematic. So I would try to improve the mesh around the sharp gradients and think whether the initial state could be made smoother.
-Peter
Re: Casting solidification
Thanks for fast ansver!