Dendritic growth

dendritic solidification, eutectics, peritectics,....
Post Reply
vronti

Dendritic growth

Post by vronti » Mon Jun 03, 2019 6:46 pm

Hello
I am relatively new to MICRESS and had been trying to learn from the forums and the documentation.
I have this assignment to grow dendrites from a wall as in a general solidification process. I have checked out T003-AlCu and T044-Dendrite model for this and tried to integrate those to create my case Al-6.2wt%Zn, for which I generated a GEN5 file with LIQUID, FCC_A1, and HCP_A3. However, for this composition only 1 solid phase exists and that way I coded my driving file.
But I am getting this weird free growth (plus shaped) dendrites instead of the directional growth from the walls.

P.S.Driving file is attached.
AlZn.dri
(14.46 KiB) Downloaded 246 times
Please help, I need to do this so badly.

Thanks,
Sreekar

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Dendritic growth

Post by Bernd » Mon Jun 03, 2019 7:20 pm

Dear Sreekar,

Welcome to the MICRESS Forum!

I think, the major problem is that you tried to define a temperature gradient in x direction. However, in MICRESS temperature gradients per default always go in z-direction. This means that you should put your dendrite on the bottom side and expect it to grow upwards.

Best wishes and good luck with your simulations

Bernd

vronti

Re: Dendritic growth

Post by vronti » Tue Jun 04, 2019 7:33 pm

Thanks for the reply,
I just tried running with the z-axis gradient. Even then it is not working and stopping at the interface mobility.

How can grow dendrites without putting nucleation sites?

Every time, I try by putting nucleation sites in further nucleation data it grows in a free growth plus shaped manner.But i need them like in a directional solidification way.

Cheers,
Sreekar

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Dendritic growth

Post by Bernd » Wed Jun 05, 2019 1:04 pm

Dear Sreekar,

What do you mean with " stopping at the interface mobility"?

As phase-field does not do automatic nucleation, you always need to start from already existing grains ("Grain Input") or set them by nucleation during runtime ("Data for further nucleation"). The typical setup for directional solidification is that you set a sufficiently strong constant temperature gradient and cooling rate and start

a) from a number of already existing grains (e.g. "small" grains with radius 0) which are set at the distance of the expected primary spacing,

b) from a flat front (defined as rectangular grain at the bottom of the domain), or

c) from a single (small) grain at the lower right or left corner of the domain which starts to form a dendrite array by side-branching.

Case c) is documented by standard examples like T001_Delta_Gamma_dri or T003_AlCu_dri. In case of b) some type of noise may be needed to trigger the breakup of the front. In all cases is can be helpful to use the moving_frame option (see the mentioned examples).

Bernd

Post Reply