MICRESS Forum

Good morning Bernd,

I am new to MICRESS and I am currently trying to simulate solidification process after laser welding process.

What I have done is created a starting surface by two grains of large radius that I have placed far below the domain and voronoi mode has been activated. Thus I can create more or less planar interface and assign different orientations to the grains.
I am using temperature values and thermal gradients uploaded from file.

1) First question that I have is large concentration gradient in front of the shrinking grains at the beginning of the simulation. I assume that it can be due to large velocities of the process that do not allow to solute enough time to diffuse away. But I am not very sure whether this assumption is valid within simulation environment.
2) another problem that I have is solute trapping. I have tried variation of the interface thickness from 2 to 3 to decrease possibility of the trapping but still I can clearly see it after simulation has stabilized.
3) For welding solidification mobility coefficient should be relatively high, I suppose. However, whenever I am trying to put values higher then 1.5 the simulation always crashes.

I would appreciate if you could help me with these issues.

Thank you!

Dear oleg4126,

Welcome again to the MICRESS forum.

I will try to answer your questions although it is not easy without knowing more details e.g. about e.g. cooling rates and temperature gradients and about the thermodynamic data you use.

1.) When rapidly melting, the tie-line is moving from the initial equilibrium values to a second operating point because the composition of the melt decreases drastically. This creates a strong concentration gradient at the melting front. But if you use a linearized phase diagram or TQ-coupling without sufficient updating, this can easily lead to negative compositions as it is the case in your simulations for component 2. Whether this is a problem for the interpretation of your simulations depends on your specific focus.

2.) How do you recognize solute trapping? Typically, this is a mainly kinetic problem and increases with increasing interface thickness...

3.) How high your mobility should be depends on the numerical conditions. What you want to achieve is diffusion limited growth. And only in case of high resolution (interface thickness << diffusion length), this is simply achieved by chosing a sufficiently high interface mobility. In all other cases (including yours), diffusion limited growth can be achieved by calibration of the interface mobility against a high-resolution test case. This procedure is described in this thread.

Please tell me if you need more informations or have other doubts.

Best wishes
Bernd