Hello,

we calculated a liquid to solid phase transformation. A typical directional growth case. Now, we want to compare the scheil gulliver mass fraction of solid over temperature curves with the MICRESS simulation. For this, we export the phase output as a text file and use python to calculate the ratio of sum of phase 1 / sum of all points, as graphically depicted in the picture below.

Is the molar volume represented in the MICRESS phase output? If yes, what we calcuated is the molar fraction of solid, correct?

Maybe someone else has faced a similar problem and choose a different calculation route to calculate the fs-T curves. I am looking forward to your replies. Thanks.

Kind regards,

Constantin & Yongliang

## Question: phase output molar or mass fraction?

### Question: phase output molar or mass fraction?

---- Constantin

### Re: Question: phase output molar or mass fraction?

Hi Constantin & Yongliang,

You are on the right track, but you should use the fraction output instead of the .phas output. If you switch on the fraction output in Micress like:

...

# Fraction output? ('frac')

# Options: out_fraction no_out_fraction [phase number]

out_fraction

...

you will get only one .frac file which displays fraction solid. If you would specify e.g.

...

# Fraction output? ('frac')

# Options: out_fraction no_out_fraction [phase number]

out_fraction 0 1 2

...

you would get individual files for each phase: .frac0, .frac1, .frac2

The advantage of using .frac0 or .frac for your analysis is that you get the correct phase fraction also at interfaces, where .phas shows "-1" only.

When I do an analysis of phase fractions over temperature, I use a Python script which averages the phase fractions along every isothermal line (or plane in 3D) so that I get a table of average fraction versus temperature for each z-coordinate. The temperature is taken from the .temp output by the same procedure (although this is somehow 'overkill' because the temperature values are identical along the isothermal line).

Finally, I join several results from different output time steps in order to get rid of the periodic oscillations which come from the side branches of the dendrite, and plot them together using gnuplot. With "smooth acspline" I get a smooth fs-T curve after selecting a suitable weight.

This type of analysis gives good results e.g. for hot cracking or freckle risk analysis. However, you will get molar fractions and no volume fractions of the phases, which you may need to correct using the average molar volumes of the phases.

Bernd

You are on the right track, but you should use the fraction output instead of the .phas output. If you switch on the fraction output in Micress like:

...

# Fraction output? ('frac')

# Options: out_fraction no_out_fraction [phase number]

out_fraction

...

you will get only one .frac file which displays fraction solid. If you would specify e.g.

...

# Fraction output? ('frac')

# Options: out_fraction no_out_fraction [phase number]

out_fraction 0 1 2

...

you would get individual files for each phase: .frac0, .frac1, .frac2

The advantage of using .frac0 or .frac for your analysis is that you get the correct phase fraction also at interfaces, where .phas shows "-1" only.

When I do an analysis of phase fractions over temperature, I use a Python script which averages the phase fractions along every isothermal line (or plane in 3D) so that I get a table of average fraction versus temperature for each z-coordinate. The temperature is taken from the .temp output by the same procedure (although this is somehow 'overkill' because the temperature values are identical along the isothermal line).

Finally, I join several results from different output time steps in order to get rid of the periodic oscillations which come from the side branches of the dendrite, and plot them together using gnuplot. With "smooth acspline" I get a smooth fs-T curve after selecting a suitable weight.

This type of analysis gives good results e.g. for hot cracking or freckle risk analysis. However, you will get molar fractions and no volume fractions of the phases, which you may need to correct using the average molar volumes of the phases.

Bernd

### Re: Question: phase output molar or mass fraction?

Yes, we also use python to calculate the fs-T curves. Thanks for your quick and helpful reply.

---- Constantin