Hello,
I simulated solidification behaviour for a B and Cr alloyed tool steel for three different cooling rates 0.1 K/s, 1 K/s and 10 K/s. Kindly check the following image which shows the dendrite growth at 50%, 65% and 100% solidification. All these simulations are having same resolutions 10µm and 0.5µm grid spacing. Surface energy and interface mobility values are optimised in order to finish the solidification without errors. All other conditions and parameters are same for three cooling rates. I have few questions on on these results.
1) For 0.1 K/s, the dendrite arms are widespread and it is growing like a cube. Is it realistic? I expected the growth of dendrite in the same manner as it does for other cooling rate i.e. 10 K/s (at 50% solidification). For 1 K/s, dendrite arms are widespread but not as much as it does for 0.1 K/s. As cooling rate decreases, dendrite arms are spreading wide (Compare column 1 images). Why there is a difference in these dendrite growth profiles?
2) For 10 K/s, I'm satisfied with the dendrite profile but as solidification proceeds, as you can see the interface is enlarging in an odd manner by retarding the growth of hard phase (In theory, higher cooling rates depresses the growth of hard phases. In this case, the primary gamma is expected to increase its phase fraction during solidification. But here, the the interface is enlarging).This is not the case with other cooling rates 0.1 and 1 K/s.
Is Surface energy has to do with this numerical error?
Kindly write me back. Thank you
Interface expansion and dendrite widespread during the solidification of a tool steel
Interface expansion and dendrite widespread during the solidification of a tool steel
- Attachments
-
- Cr10-3b_CoolingRates_comparison.png (859.39 KiB) Viewed 1699 times
Re: Interface expansion and dendrite widespread during the solidification of a tool steel
Dear Vamsi,
I think, your results look quite reasonable. If cooling rate is reduced for a fixed domain size (i.e. PDAS), there should be a transition from dendritic to cellular, which you observe indeed.
However, this setup is a bit artificial as in reality the PDAS would not be fixed. Instead, it would increase so that the dendritic structure remains similar.
With increasing cooling rate, irrespective of the PDAS, the eutectic structures get finer and finer, and cannot be properly resolved any more. This you can see as broadening of the interfaces.
What you should do is to rescale as well the PDAS as the grid resolution with the cooling rate. Then, you will get self-similar structures, and you can really compare whether you get more or less "hard" phases. However, instead of choosing interface mobility and energy "in order to finish the solidification without errors", you should use a physically reasonable interface energy and a diffusion limited interface mobility by using the "mob_corr" option...
Bernd
I think, your results look quite reasonable. If cooling rate is reduced for a fixed domain size (i.e. PDAS), there should be a transition from dendritic to cellular, which you observe indeed.
However, this setup is a bit artificial as in reality the PDAS would not be fixed. Instead, it would increase so that the dendritic structure remains similar.
With increasing cooling rate, irrespective of the PDAS, the eutectic structures get finer and finer, and cannot be properly resolved any more. This you can see as broadening of the interfaces.
What you should do is to rescale as well the PDAS as the grid resolution with the cooling rate. Then, you will get self-similar structures, and you can really compare whether you get more or less "hard" phases. However, instead of choosing interface mobility and energy "in order to finish the solidification without errors", you should use a physically reasonable interface energy and a diffusion limited interface mobility by using the "mob_corr" option...
Bernd