Epitaxial growth in AM

dendritic solidification, eutectics, peritectics,....
Post Reply
Moritz
Posts: 40
Joined: Fri Aug 06, 2021 11:13 am
anti_bot: 333

Epitaxial growth in AM

Post by Moritz » Thu Jan 06, 2022 12:28 pm

Hello everyone!

I hope everyone had a great start into 2022!

I am still working on a model for the solidification in the LPBF process. As the model develops, I want to include one more significant characteristic of real Am processes, which is epitaxial grain growth. What I mean by that is the orientation of the grain growth towards a certain direction in a certain angle. This is shown in the paper 'Understanding grain evolution in additive manufacturing through modeling' by Akram et al. 2018 from in figure 1:

Image

To achieve this, I thought the following parameters set in the driving file were sufficient:

Image

The initial grains at the bottom look like this, after a few seconds into the simulation:

Image

Unfortunately, after the solidification has gone on for more steps, no orientation of the forming microstructure can be observed:

Image

This is also the case, if the simulation has finished. Between a simulation with specified grain orientations and without there is no difference.

Has anyone a suggestion, what I am missing here?

I am thankful for every reply.

Best regards,

Moritz

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

Re: Epitaxial growth in AM

Post by Bernd » Thu Jan 06, 2022 5:36 pm

Hi Moritz,

wish you also a Happy New Year.

Apart from defining orientations for the initial grains, you also need to introduce anisotropy to the phase interaction (interface mobility and interface stiffness). Otherwise, orientation would not have any effect on growth.

Please also take care that grid resolution is sufficiently fine. Otherwise, numerical grid anisotropy will be dominating and make the dendrites grow along the z-direction. Furthermore, a too coarse grid resolution will lead to a high kinetic undercooling (when using "mob_corr" for diffusion limited kinetics) which makes effects of curvature and anisotropy waning (see e.g. my comments here).

Bernd

cboe
Posts: 14
Joined: Thu Feb 01, 2018 6:06 pm
anti_bot: 333

Re: Epitaxial growth in AM

Post by cboe » Fri Jan 07, 2022 1:50 pm

Hi Moritz,

I thought about simulating a similar scenario for laser beam welding in a 2d heat conduction case. I have not come up with a working simulation yet. Maybe my humble considerations will help you.

As far as I know the angle of the dendrite has two sources i) the orientation of the parent grain/dendrite and ii) the temperature gradient. From welding metallurgy I know that the latter ii) is more important. Therefore, the direction of the dendritic growth is perpendicular to the local liquidus isotherm surface (3d) or line (2d). However, this angle changes from the beginning to the end of the solidification. In the begining (at the maximum weld pool width which is the starting point of solidificaiton) the angle between laser direction and the normal to the liquidus isotherm is 90° -> no growth occurs. Afterwards, it changes from 90° to 0° (in the weld pool centre where - in a steady-state condition - the laser velocity is equal to the solidification rate).

When thinking about solidificaiton in laser beam welding or LPBF, I arrange my simulation box also perpendicular to the local liquidus surface/line because MICRESS is limited to only a temperature gradient in the z-direction. However, in a realistic simulation the direction of the gradient changes. So either the box needs to rotate according to the local angle OR the temperature gradient needs to change.

Not sure if this helps or confuses more :)

Best
Constantin
---- Constantin

Moritz
Posts: 40
Joined: Fri Aug 06, 2021 11:13 am
anti_bot: 333

Re: Epitaxial growth in AM

Post by Moritz » Wed Jan 12, 2022 9:26 am

Hello everyone,

thank you for your replies and suggestions. I am going to implement them into future attempts.

I also already thought about the orientation of the heat source and how I could resemble the temperature conditions. I think it is an interesting idea to turn the simulation, to overcome the limitation of the temperature gradients in z-direction.

Best regards,
Moritz

Post Reply