Page 6 of 6

Re: Primary Dendrite Arms spacing (PDAS) prediction

Posted: Wed May 23, 2018 12:11 am
by Bernd
Dear Shenyz,

I think there are several issues which lead to very slim dendrites:

1.) The initial undercooling is much too high. You should start at a 20K higher temperature...

2.) The interface mobility is about a factor 2 too high. I found that using "mob_corr" in the new version 6.4. Depending on your MICRESS version you should either use mob_corr or reduce the interface mobility.

3.) Due to a relatively low resolution, the interface gets unstable despite of the high averaging value of 1. The interface spreads in grid direction which leads to a too low tip undercooling and to too slim dendrites. The solution is either to increase resolution (which costs performance) or to introduce an extra interface stabilisation (second optional parameter after interface energy with about 10 times the value of the interface energy). The stabilisation term prevents interface spreading without affecting the physical value of the interface energy, and a more realistic dendrite shape is obtained without needing to increase resolution (look here).

Apart from that, you loose a lot of performance with updating linearisation data in each interface cell. Using "global" relinearisation mode would speed up your simulation tremendously...


Re: Primary Dendrite Arms spacing (PDAS) prediction

Posted: Wed May 23, 2018 6:11 am
by Shenyz
Dear Bernd:

Thank you very much for your advice and I will modify my project. Hope to get good results.