Cr-Ni eutectic alloy simulation

dendritic solidification, eutectics, peritectics,....
lihaoge
Posts: 22
Joined: Mon Nov 06, 2017 3:16 am
anti_bot: 333

Cr-Ni eutectic alloy simulation

Post by lihaoge » Mon Nov 06, 2017 3:44 am

Hello
I am a new learner of Micress and hope someone can help me!
I want to simulate the eutectic process of Cr-Ni alloy, so I read the driving file from example "TQ_Eutectic_in.txt" and made some changes on this basis and created the GES5-File from database TCFE7 and MOB2. I have a diffusion database MOBFE2, how to use these data to specify diffusion data like Diff.-coefficient Prefactor for diffusion of element 1 in phase 0 or 1 in phase 1 or its activation energy?
Thank yo very much!

Bernd
Posts: 1074
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cr-Ni eutectic alloy simulation

Post by Bernd » Mon Nov 06, 2017 11:34 am

Dear lihaoge,

Welcome to the MICRESS forum.

For using diffusion data from database in MICRESS, mobility data must be added to the thermodynamic data in the .ges5 file. How this is done is demonstrated in the Thermo-Calc macro file "HOWTO_FeCMn.TCM" which is included in the redistribution examples and which you can use as a template.

It is important to consider first which mobility database fits to which thermodynamic database. This is essential for two reasons:
- First, in order to calculate diffusion coefficients from the mobilities in the database, thermodynamic data (thermodynamic factors) are needed. Inversely, for fitting mobility data during assessment using experimental diffusivity data, thermodynamic data have been used. Thus, if mobility data is used together with other thermodynamic data than those used during assessment, wrong diffusivities may be calculated which do not represent the underlying experimental data.
- Secondly, it is technically required that not only the phase names but also the phase models (number of sublattices, etc.) in the thermodynamic and mobility database fit each other.

When simulating iron based materials, the mobility database
MOB2 should be used together with thermodynamic databases SSOL and TCFEx (x<=5), while
MOBFEx (always newest, i.e. highest x available) should be used for all the newer TCFEx (TCFE6 and newer)
This is based on information on the Thermo-Calc webpage. Corresponding comments in the "HOWTO_FeCMn.TCM" file may be slightly outdated, we are sorry for that...

Practically that means for you to create a .ges5 file on the basis of TCFE7 and MOBFE2 using a modified "HOWTO_FeCMn.TCM" script.

In the MICRESS input file you specify diffusion data in the section "# Diffusion Data". If you want to use only diagonal diffusion terms (like you should do for the liquid phase), the keyword is "diagonal g" (or "diagonal l" in case you further want to include interpolation of concentration dependency). For multicomponent diffusion (which includes off-diagonal terms) you use "multi" followed by as many characters as you have components (e.g. "multi ggg" if your number of dissolved elements is 3).

Bernd

lihaoge
Posts: 22
Joined: Mon Nov 06, 2017 3:16 am
anti_bot: 333

Re: Cr-Ni eutectic alloy simulation

Post by lihaoge » Mon Nov 06, 2017 1:57 pm

Dear Bernd,

Thanks for the warm welcome and the comprehensive reply. I created the .ges5 file and start to calculate. However, I found the results that only shows process of a grain growth, not the eutectic process. Maybe I have some problems when defining the driving file. Do you have any ideas what reason lead to this result?

Thank you very much.
Attachments
CrNi_Eutectic_korn_32.png
CrNi_Eutectic_korn_32.png (12.43 KiB) Viewed 1359 times
CrNi_dri.txt
(22.73 KiB) Downloaded 69 times

Bernd
Posts: 1074
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cr-Ni eutectic alloy simulation

Post by Bernd » Mon Nov 06, 2017 3:32 pm

Dear lihaoge,

the reason for that is that you obviously define an initial microstructure ("grain input") which you have taken from another simulation where a single phase grain structure was required. Especially, the "voronoi" keyword makes sure that no liquid phase (phase 0) remains between the grains of phase 1.

I don't know how you wish the initial microstructure to look like - probably it will depend on the type of calculation you want to do. Is it a directional growth (lamellar eutectic which grows along a temperature gradient) or rather equiaxed solidification with interdendritic (or intercellular) eutectic phase?

Bernd

Bernd
Posts: 1074
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cr-Ni eutectic alloy simulation

Post by Bernd » Mon Nov 06, 2017 3:36 pm

By the way, I put some more general information about how to define diffusion data here (although "multi_plus" is not available before MICRESS version 6.4).

Bernd

lihaoge
Posts: 22
Joined: Mon Nov 06, 2017 3:16 am
anti_bot: 333

Re: Cr-Ni eutectic alloy simulation

Post by lihaoge » Tue Nov 07, 2017 4:05 am

Dear Bernd,

Thanks for the comprehensive reply. The initial microstructure is a directional growth (lamellar eutectic which grows along a temperature gradient). Do you have any Idea that how to define the "grain input"?
And I defined the diffusion data like:

# How shall diffusion of component 1 in phase 0 be solved?
diagonal d
# Diff.-coefficient: (in case you prefer another value than 1.E-5cm2/sec)
# Prefactor? (real) [cm**2/s]
2.00000E-05
# Activation energy? (real) [J/mol]
0.0000
# How shall diffusion of component 1 in phase 1 be solved?
diagonal g
# How shall diffusion of component 1 in phase 2 be solved?
diagonal g
# How shall diffusion of component 2 in phase 0 be solved?
diagonal d
# Diff.-coefficient: (in case you prefer another value than 1.E-5cm2/sec)
# Prefactor? (real) [cm**2/s]
2.00000E-05
# Activation energy? (real) [J/mol]
0.0000
# How shall diffusion of component 2 in phase 1 be solved?
diagonal g
# How shall diffusion of component 2 in phase 2 be solved?
diagonal g

But the MICRESS got error:

In leseConc answer = diagonal d
Input error in routine leseConc

Do you have any Idea what these errors are?

Thank you very much!

Bernd
Posts: 1074
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cr-Ni eutectic alloy simulation

Post by Bernd » Tue Nov 07, 2017 12:44 pm

Dear lihaoge,

The diffusion input itself seems correct. Probably, the error occurred before, and MICRESS expected another input at this place. You can find it out if you compare the screen output with the order of input in the input file. Which input was requested when the wrong answer came? Sometimes the error happens even several input lines before it stops, and MICRESS already accepted some wrong input lines because of correct format if they were erroneously shifted...

For grain input, in case of directional lamellar growth, I would define initial grains only at the bottom - either of both phases or only of one with later nucleation of the second. Furthermore, you can either define the initial grains "deterministic" (like in the "Delta_Gamma_dri" example), which means you specify coordinates and radius of each of them individually, or "random" (like you did). In the latter case, you should restrict the z-range to a smaller area at the bottom.

Bernd

lihaoge
Posts: 22
Joined: Mon Nov 06, 2017 3:16 am
anti_bot: 333

Re: Cr-Ni eutectic alloy simulation

Post by lihaoge » Wed Nov 08, 2017 12:15 pm

Dear Bernd,

Thank you very much! I used the grain input data from "Delta_Gamma_dri" example and see what happen. But MICRESS display that:

toolGetEnthalpy: Error1 1607
Thermo-Calc error 1607 MICRESS error 24 phases 0/0

*** ERROR 1607 IN QEQUIL
*** DEGREES OF FREEDOM NOT EQUAL TO ZERO
Thermo-Calc error 1122 MICRESS error 30 Thermo-Calc

If something went wrong when creating the .GES5 file, then why it would not happen before I change the grain input data?
Do you have any ideas?

Thanks!
Attachments
CrNi_Eutectic_log.txt
(13.37 KiB) Downloaded 63 times
CrNi2.txt
(22.16 KiB) Downloaded 61 times

Bernd
Posts: 1074
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cr-Ni eutectic alloy simulation

Post by Bernd » Wed Nov 08, 2017 1:21 pm

Hi lihaoge,

The difference is that now with the changed grain input your domain has liquid phase - and the error occurs when calculating the enthalpy of phase 0 (liquid)!

Thus, if there is something wrong with the .ges5 file, then it is associated to the liquid phase. To avoid incompatibilities between different versions, please try to use the newest MICRESS version 6.300. Perhaps it helps if you avoid to include mobility data for the liquid phase into the .ges5 file.

Apart from that, you should not use latent heat (and calculate enthalpy values) in case of directional growth. In the presence of strong temperature gradients, latent heat release can be easily neglected, and a cooling rate and temperature gradient should be defined instead. Just change "lat_heat" for "no_lat_heat" and change from heat extraction rate to cooling rate and define your temperature gradient in the "Boundary Conditions" section.

Bernd

lihaoge
Posts: 22
Joined: Mon Nov 06, 2017 3:16 am
anti_bot: 333

Re: Cr-Ni eutectic alloy simulation

Post by lihaoge » Thu Nov 09, 2017 4:10 am

Dear Bernd,

Thanks for your reply! That error do not happen after I avoid to include mobility data for the liquid phase into the .ges5 file. And I changed the Boundary Conditions as you said. But there is no calculation results finally. Do you know why this happen?

Thanks! 
Attachments
CrNi2.txt
(22.61 KiB) Downloaded 63 times
1.png
1.png (10.82 KiB) Viewed 1321 times

Post Reply