using appropriate phase to mimic substrate layer in LPBF process

[Hanyu]

Hello everybody,

I select X20MnAl12-3 as my simulated materials to simulate the solidification process under additive manufacturing conditions. I use Rosenthal solution to generate a temperature profile. The primary phase is ferrite (white color), followed by the growth of austenite (orange color) in the interdendritic region of the ferrite dendrites. The results obtained when using the austenite phase to mimic the substrate layer are shown

Screenshot 2024-02-02 203309.png (402.66 KiB) Viewed 266 times

The large size austenite grains grow in front of the ferrite dendrites by the end of simulation.

However, if the ferrite phase is selected to mimic the substrate layer, the results appear as shown

Screenshot 2024-02-02 204421.png (138.28 KiB) Viewed 266 times

Both simulations use the same temperature file, but the substrate phase differs. The second simulation is considered more realistic. Could you please explain the reason for the appearance of the large size austenite grains in the first simulation results? Thank you in advance.

temperature profle.txt

(10.07 KiB) Downloaded 29 times

first simulation code.dri

(40.52 KiB) Downloaded 32 times

second simulation code.dri

(40.97 KiB) Downloaded 28 times

TC2022b_FeCMnAl.TCM

(515 Bytes) Downloaded 25 times

TC2022b_FeCMnAl.GES5

(167.04 KiB) Downloaded 26 times

[Bernd]

Dear Hanyu,

Welcome to the MICRESS Forum! You describe a case where 2 phases can equally well grow as primary phase and thus compete with each other. This means, the question which one will prevail or overgrow the other strongly depends on the nucleation conditions for each phase. And the opportunities for nucleation also depend on the process dynamics and thus on the initial state.

However, the question is whether or how much we can trust the outcome of the simulation in such cases. Often we do not really know much about how nucleation proceeds in reality, and then with simulation we only can compare different scenarios (assumptions), which we have to check against reality.

In your case, it is the initial microstructure which makes the difference: In your second simulation, you start with a single crystal of phase 2 which has optimal orientation (0°) with respect to the temperature gradient. Thus, a perfect dendrite front is formed which grows at high temperature and does not have grain boundaries. Contrarily, in your first case, you have random nucleation of phase 2, and the dendrites grow with less favorable directions and also with grain boundaries, making it much easier for phase 1 to step in by nucleation. For example, by starting your second case not with a single but with various grains of phase 2 with different orientations you could check whether my suspicion is correct or not...

Bernd