Primary Dendrite Arms spacing (PDAS) prediction

[Bernd]

Hi Chamara,

As you can see from the output which you sent, temperature is already around 600K after 1 second, so you should not expect that you get anything reasonable for the liquid/fcc_L12 interface

So, the basic problem is that your domain size, the cooling rate and the interface mobility are not yet in reasonable relation.

I would further suggest to use automatic mobility correction ("mob_corr") which you can define in the phase diagram data section for each element once you have chosen "redistribution_control" in the phase interaction data. This gives you at least an idea about reasonable interface mobility ranges.

Bernd

[CharMIC]

Hej Bernd,

Many Thanks for the advice.

Do I also have to adjust the numerical time step due to large cooling rates and thermal gradients to fit the growth velocities?

BR Chamara

[CharMIC]

Hej Bernd,

Addition to that is there any way to get these interface mobility data from Thermo-calc data base.

BR
Chamara

[Bernd]

Hi Chamara,

The numerical time step of MICRESS is typically chosen automatically (given that you specified "automatic" in the time input data. Then, there is no need to change that.
However, in case of "automatic_limited", a minimum time step may have been specified to improve performance. This should always be made very small, or "automatic" should rather be chosen when conditions change. Finding a new value for the time step limits as part of optimization of the new simulation once it runs stable.

The interface mobility is not stored in thermodynamic databases. Furthermore, in our case of probably diffusion limited growth, it is also not a physical parameter but only a numerical one which should have the correct value to give correct interface kinetics. It will change always if you change spatial resolution or alloy composition.

Bernd

[CharMIC]

Hej Bernd,

Thanks for the explanation.

As you said, is the interface mobility only change when you change spatial resolution or alloy composition.

I am asking this since if I calibrate the interface mobility, so that I get the correct PDAS for given cooling rate and thermal gradient, Can I use the same calibrated value for different cooling rates and thermal gradient to predict PDAS (assuming I don't change composition and spatial resolution)?

BR
Chamara

[Bernd]

Dear Chamara,

In first approximation, yes. However, you can change these parameters only in a relatively small range without needing to change resolution...

Anyway, I above recommended to use "mob_corr", then you do not need to worry about calibration of mobility!

The option has been greatly improved in the upcoming version 6.4 but should already work with MICRESS 6.3 for your application.

Bernd

[CharMIC]

Hej Bernd,

Regarding the mob_corr...

I cannot find description in the manual for the mob_curr option.

Could you explain little about it and how it work.

Which mob_corr you advice to use. Normal or the atc ?

BR
chamara

[Bernd]

Dear Chamara,

in short: You need to set the "redistribution_control" option in the corresponding phase interaction, and put for each element in the phase diagram data the line

normal mob_corr

The mobility value still needed in the phase interaction data should be higher than what you expect to come out from "mob_corr" (e.g. 100.). You can check the .mueS output to see the calculated mobility.

I plan to give a more detailed description in this forum soon.

Bernd

[Bernd]

I have put some more information there

http://board.micress.de/viewtopic.php?f=26&t=527

Bernd

[CharMIC]

Hej Bernd,

Thanks for your valuable help.

I was wondering what is your experience on the mobility value range for the Liquid/gamma phase interaction in Nickel based super alloys. I saw there are some example pictures of simulating dendrite growth of IN718 and IN706.

I just want to know to have a feeling of the possible value or range?

BR
Chamara