MICRESS Forum

Hi,

To define a particular geometrical shape, I had to include a grain which I do not want to be interacted with adjacent or any other phases.
for example,

-phases in the domain are liq,bcc,fcc and diamond and I dont want diamond to have any interaction with any other phases.
-diamond and bcc have a common grain boudery
-defined phase interaction in diamond only with liquid. (liq does not come in-contact with dia whatsoever)
-no diffusion in dia phase.

Question;
1.composition of the all phases at t=0, were 0.18 (wt%). at the common grain boundary, dia composition went down to negative 12!!!. (-12)(wt%). did I read it wrong or is it x*10^-12 ?
2.does this have any effect on the average compositions in liq,bcc and fcc.

Could you please let me know your thoughts in it, advice me if you need the ciode.
Thank you

Regards,
Dasith

Dear Dasith,

Welcome to the MICRESS forum!

In such a case, you should not define an interaction between liquid and diamond.

MICRESS cannot know that these two phases never will get into contact. Typically, undefined phase interactions are completed at initialisation time (or during runtime in triple junctions) based on the existing interactions. For example, if you would have two phases in contact with liquid, and no phase interaction between the two solid phases defined, the phase diagram between the solids would be estimated from the two solid/liquid interactions in such a way that a constant partition coefficient is assumed ("constant-K approximation"). In case of local phase diagram descriptions, this is done for each interface cell in the triple junctions. This is important if diffusion between the phases has to be calculated. However, if there is not sufficient information like in your case, this approximation can fail.

Things are different if for a phase there is no interaction at all. Then, the phase is assumed to be "inert", and no interactions are calculated at all. This is what you should do!

Bernd

Dear Bernd,

Thank you very much for your descriptive answer, it is very clear now.

I have an another question.
Temperature gradient is defined alone the z-axis.

When we define the cooling rate is it defined on the same z-axis? if yes, can we alter it?

Thank you very much
Kind Regards,
Dasith

Dear Dasith,
Temperature gradient is defined within the simulation domain(temperature change in space), whereas the cooling rate definition is for the cooling rate to be followed with time (temperature change with time).

Best Wishes,
Deepu