Analyzing average compositions in dendritic and interdendritic regions

aspects of evaluating simulation results and their graphic presentation using either DisplayMICRESS or other software tools. Features and possibilities of DisplayMICRESS
Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Analyzing average compositions in dendritic and interdendritic regions

Post by Bernd » Tue Feb 08, 2022 7:16 pm

Dear Ahmet,

First of all, I think it is dangerous to base your observations on the peak concentration. Both, in experiments as in simulation, if you have Al-rich precipitates, the peak concentration depends on the discretization size (beam size, grid size) and the relative particle sizes. Instead, you should definitively use some integral measures for your comparison.

I can't follow your argument, that with fast cooling segregation should be less strong because "the diffusion distances is shorter", and "therefore there might be no time to re-arrange this concentration spike/gradient". The reason is that the microstructure length scale (primary/secondary spacing) is also correspondingly smaller, so that from this point of view there should be no difference. You can only have a difference from this source, if the morphology is not scaling with the diffusion length, i.e there is a systematic change or morphology, e.g. from more dendritic to more cellular.

A contribution which does not scale in this way is curvature and the curvature undercooling. There must be a systematic shift of the solidification temperature to lower values with increasing cooling rate. This temperature shift can affect the solubility of Al in the primary phase, leading to more or less strong segregation. If you want to evaluate this effect with Thermo-Calc, you would have to keep the phase fractions constant while lowering the temperature (so that the resulting driving force resembles the effect of curvature on the equilibrium compositions).

Another quantity which also does not scale with the cooling conditions is the physical interface thickness (which we assume to stay the same) and which causes solute trapping. This effect is not automatically included in MICRESS (essentially it is even prevented by using the thin interface correction models). However, I believe that solute trapping is still not relevant under all three condition which you compare.

I agree fully to your statement that the homogenization time strongly decreases for higher cooling rate microstructures, so that even the relatively short reheating or elevated process temperature can be responsible for reduced segregation found in the experiments in case of SLM.

Bernd

Atur
Posts: 33
Joined: Tue Apr 13, 2021 11:39 am
anti_bot: 333

Re: Analyzing average compositions in dendritic and interdendritic regions

Post by Atur » Wed Feb 09, 2022 3:56 pm

Dear Bernd,

Thank you for your reply.

You are right, It was silly that I did not consider the grid size effect, interface thickness and undercooling between these comparisons. Indeed they are adjusted differently in different processes. Yet I observe the same results when I compare the setups with same interface thickness & grid size. But I assume then it is also because of highly resolved precipitates in fine grids.

I am just trying a basic and fair set-up to extract segregation related effects, and how do they evolve over a given time (specifically right after solidification). But so far I could not go further than analyzing the relative fraction of segregated area to the total area, which is anyway an obvious and expected outcome. Are there any other relations you can suggest me to scale and compare trends of concentration gradients in interdendritic regions? I am afraid I could not discover a clear rule how to do it for obtained representative microstructures.



regards,
Ahmet

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

Re: Analyzing average compositions in dendritic and interdendritic regions

Post by Bernd » Wed Feb 09, 2022 9:06 pm

Dear Ahmet,

As I already mentioned, by comparing integrals of the concentration deviation from the average over a certain part of the concentration distribution (as e.g. obtained by the DP_MICRESS segregation analysis) it should be possible to get an analysis method which should not be sensitive to unresolved precititates. You could e.g pick 1/3 of the total area with the highest .concAl values and integrate (c - c_0)/c_0. This would give you sort of a dimensionless segregation coefficient which you could use for comparing different simulations for different conditions. Ideally, this coefficient should not change for a given setup if you only change numerical parameters like grid resolution.

However, when comparing to experimental concentration mappings, there is a systematic deviation which is caused by the 2D-nature of the simulations...

Bernd

Post Reply