fix/update Marangoni coefficients in Mist files

[gknapp1]

• Fixes the sign in the IN625 & IN718 Mist files to match the values in the reference
• Adds Marangoni coefficients to the SS316L and SS316H file, along with a note about the assumptions of the value (0 ppm Sulfur at 1925 K)
• These values are not used anywhere in Myna at the moment, so this doesn't change behavior of any existing functions

Closes #132

[colemanjs]

Your values are all 1e-5 N/(m-K), but should be 1e-4 N/(m-K), correct?

[gknapp1]

Your values are all 1e-5 N/(m-K), but should be 1e-4 N/(m-K), correct?

The -2e-5 values for IN625 & IN718 are correct (or at least agrees with the Gan et al. values that are cited). But this was a typo in the SS316 values. Fixed.

[colemanjs]

Your values are all 1e-5 N/(m-K), but should be 1e-4 N/(m-K), correct?

The -2e-5 values for IN625 & IN718 are correct (or at least agrees with the Gan et al. values that are cited). But this was a typo in the SS316 values. Fixed.

Interesting. Digging further, the -2e-5 value reported in Gan et al. is a calibrated value for their in-house thermofluid model to match single track melt pool geometries from NIST. I guess this will depend on how we want to define the MIST properties: Should they be effective or calibrated values for modeling (in which case we need to make sure all other properties and problem definitions are the same between the source and our application), or do we want to prioritize measured properties from experiments and CALPHAD whenever available?

[gknapp1]

Interesting. Digging further, the -2e-5 value reported in Gan et al. is a calibrated value for their in-house thermofluid model to match single track melt pool geometries from NIST. I guess this will depend on how we want to define the MIST properties: Should they be effective or calibrated values for modeling (in which case we need to make sure all other properties and problem definitions are the same between the source and our application), or do we want to prioritize measured properties from experiments and CALPHAD whenever available?

For the MIST data in Myna, I think it should represent our "best practices" for material data. In some cases, this will be material property values that we have validated thoroughly, but I think in other cases it will be a best guess based on uncertain measurements and simulated values that worked in other research papers.

I did add a note in the Inconel files to highlight that the values were calibrated. I knew that, but it is a good thing to expose in the MIST file. I also opened ORNL-MDF/mist#33 to make assumptions easier to document in the future.