MICRESS Forum

Hi,

I am doing a simulation of dendrite growth without any thermal gradients. I have to seeds type at the bottom of the domain. in the right half I have seeds with an orientation of +8 degrees and in the left half I have -8 degrees. However when the dendrites grows, it initially growth parallel to the angles that I have given (+8 and -8) however, after a certain point it align it self parallel to the Z direction. I do not have a thermal gradient in the simulation. Only a cooling rate. How can I mitigate this issue?

BR
Chamara

Dear Chamara,

My understanding of this effect is that there is an instability with slight broadening of the interface which starts when the dendrite growth speeds up. This "spreading" or broadening is easier in grid direction.
In principle, all measures which stabilise the interface profile should help to prevent that. If you are not able to increase resolution, I would try to use interface stabilisation (second optional parameter after interface energy). Using random noise on the averaging direction of the driving force ("smooth") could also be helpful.

Bernd

HI Bernd

how does this maximum value for the interface stabilization works?

BR
Chamara

Hi Chamara,

The principle of the stabilisation term is quite simple: A difference term is obtained by subtracting curvature from the interface term K_alphabeta (as defined here). This stabilisation term can be scaled (by a certain extent) without creating artificial curvature effects. Afterwards, curvature is added back.
The procedure has been generally described in Y. Sun, C. Beckermann, Journal of Computational Physics 220 (2007) 626–653 (especially Eq.(12)). There is no publication on the specific implementation in MICRESS.

Bernd