Solidification for LPBF-AlMg alloys

dendritic solidification, eutectics, peritectics,....
Post Reply
Desai
Posts: 5
Joined: Fri Oct 15, 2021 3:03 pm
anti_bot: 333

Solidification for LPBF-AlMg alloys

Post by Desai » Wed May 18, 2022 1:51 pm

Dear All, Good day.

I am working on the solidification simulation of Al-Mg alloy in LPBF and then want to make a comparison of the simulation microstructure with EBSD results.
I have selected the phases and their interactions based on the Scheil diagram (FIGURE A).
For #Process_conditions I have used Time-Temperature data.
The position of equiaxed and columnar grains in the simulation is as expected.
I have got the following results so far (FIGURE A):
.
.
SCA image 1.jpg
SCA image 1.jpg (218.21 KiB) Viewed 3941 times
.
.
However;
1) In the initial Microstructure the #Phase_number_for_grain_type was assigned as LIQUID but no results were obtained (FIGURE B).
On changing the #Phase_number_for_grain_type as FCC_A1, the results FIGURE A was achieved.
However, is this a right approach to assign the phase for grains as FCC_A1 and not as LIQUID in the Initial Microstructure?
SCA 2.jpg
SCA 2.jpg (36.6 KiB) Viewed 3941 times
.
.
2) The effect of Nucleation is NOT observed (Seed type for all the phases according to Scheil were selected).
Even on choosing #NO_NUCLEATION input, the same results were obtained as in (FIGURE A)
Is it because of the defined Initial Microstructure?
How do I address this issue of Nucleation.

3) Is it possible to simulate and view the formation of Precipitates in the melt and
their interactions which lead to the formation of equiaxed grains?
.
.




.
.
I am very thankful for the suggestions.
Best Regards.
Last edited by Desai on Fri May 20, 2022 6:45 pm, edited 1 time in total.

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

Re: Solidification for LPBF-AlMg alloys

Post by Bernd » Wed May 18, 2022 8:28 pm

Dear Desai,

here are some answers to your questions:

1. Initial Microstructure
Setting initial grains with phase 0 (liquid) has not much effect because the default state of the initial microstructure is phase 0 everywhere. If you want to start from pure liquid phase, just do not set any initial grains. But, as nucleation is not working currently (see below), nothing would happen afterwards.
However, given the application case of additive manufacturing and the initial temperature profile which you are reading in, I suspect you want to start from an initial microstructure which is solid at the bottom (dense multi-grain structure of fcc) and fully liquid at the top. To achieve that you should
- remove all but one of the initial grain types (I don't see why you should need more than one ...)
- apply "voronoi" construction in order to get a dense (completely solid) fcc-structure
- adjust the maximum z-range to a value which corresponds to the upper end of the solid region. It makes sense to adjust the first temperature profile such that at the upper end of solid the initial temperature is close to the liquidus temperature.

2.) Nucleation input data must be adjusted to the correct time- and length scales. If you are unsure about nucleation setiings, you should activate the "verbose" switch so that you get information about each time when nucleation is checked.
- decrease the time interval for checking each nucleation type to a much smaller value. Currently you barely check even once during the whole simulation time!
- you should not use a shield distance of 0. Use a realistic value instead (or better activate the second optional input of the nucleation distance, which allows to control how much is the distance between checked points). Otherwise, checking will be extremely slow because too many points are checked, and you will get many seeds at once.
- for eutectic phases I would recommend using "interface" nucleation.

3.) With the above recommendations for the nucleation input data you should be able to simulate and view the formation of precipitates in the melt and their interactions which lead to the formation of equiaxed grains.

Bernd

PS: For simulation of additive processes I strongly recommend using our latest version 7.1 which is specifically dedicated to additive processes. It contains a bunch of functionalities that are crucial for situations where strong temperature gradients lead to big temperature differences within the simulation domain.

Desai
Posts: 5
Joined: Fri Oct 15, 2021 3:03 pm
anti_bot: 333

Re: Solidification for LPBF-AlMg alloys

Post by Desai » Wed May 25, 2022 2:45 pm

Dear Bernd,

Greetings for the day :) . Thank you very much for your quick reply.
I have applied your suggestions to the simulation regarding the 'Initial microstructure' and also in the 'Nucleation' section.
and have used Nucleation distance input instead of the shield distance.
I have got the following results.
.
.
download.jpg
download.jpg (173.53 KiB) Viewed 3905 times
download2.jpg
download2.jpg (125.28 KiB) Viewed 3905 times
.
.
.
However,
1) After certain iterations there was an error message 'ERROR IN FIRST EQUILIBRIUM: AL3X'.
How do I fix this issue.
.
Error micress.jpg
Error micress.jpg (443.72 KiB) Viewed 3905 times
.
.
2) Also the microstructure now only has Equiaxed grains formation, What causes that there is no observation of columnar grain formation??
Thank you :) .
.
.
Last edited by Desai on Sat May 28, 2022 7:23 pm, edited 1 time in total.

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

Re: Solidification for LPBF-AlMg alloys

Post by Bernd » Wed May 25, 2022 8:03 pm

Dear Desai,

The error which you experienced happens when trying to calculate the molar volume of the Al3X-phase. One reason could be that there is a stoichiometric condition missing: Depending on the phase constitution, MICRESS sometimes cannot automatically detect stoichiometric phases, e.g. if the database description of the phase contains species which are not elements. Then, you must manually define the phase as stoichiometric for all element without solubility range (in other cases it also does not hurt...). Whether a phase is stoichiometric or not can be seen in the "Start Compositions and Limits for Quasi-Equilibrium calculation" output which is written to the .log-file and on screen output after initialisation.

From your simulation results it seems that you already have an undercooled melt region above the initial solid front at t=0, which does not change much during the following time steps. This "quenching always will lead to equiaxed growth. It would be better to start with an initial temperature profile which has the liquidus temperature close to the initial front position.

Bernd

Desai
Posts: 5
Joined: Fri Oct 15, 2021 3:03 pm
anti_bot: 333

Re: Solidification for LPBF-AlMg alloys

Post by Desai » Wed Jun 01, 2022 2:11 am

Dear Bernd,

I tried the simulations after applying Stoichiometric conditions.
However, I still have the same Error.
The phase Al3X is not present in the Mobility database of the GES file, Could this also be a reason for this error?

stoichio.png
stoichio.png (166.21 KiB) Viewed 3865 times
.
.
Last edited by Desai on Tue Jun 21, 2022 12:14 pm, edited 1 time in total.

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

Re: Solidification for LPBF-AlMg alloys

Post by Bernd » Fri Jun 03, 2022 7:42 pm

Hi Desai,

No, it has nothing to do with diffusion data, the error occurred while calculating volumetric data.

I tried to reproduce the error, but could not. Instead, there are thousands of other error for interface 0/4 and also for 0/2. My impression is that the basic problem is a too coarse grid resolution, leading to high values of the driving force for the primary phase (interface 0/4, in connection with mob_corr), and thus also to extremely metastable equilibria. So, at the moment, I would suppose that the error which you mentioned is just a secondary problem...

My suggestion is to try higher resolution, first starting from a very small domain. Then you can see whether the problems vanish.

Bernd

Post Reply