Updating of diffusion data from database...

[SuperX]

Hello Bernd:

I am using Micress to simulate the solidification of ductile cast iron. A problem occurs when I run my driving file:
When the first seed occurs in the bulk, the simulation stops, with the screen displaying such a sentence: Updating of diffusion data from database...
I have used a database before, but I don not know where the problem is.

Hope for your reply

Best Regards

[Bernd]

Dear SuperX,

Welcome to the MICRESS forum.

I guess when first nucleation happens, it will be GRAPHITE which is formed. Then, diffusion data are requested from the database. Are you sure that there is diffusion data available there? I doubt that there is diffusion data available. Even if it would, it does not make sense to calculate diffusion of C in stoichiometric pure C (graphite).
Anyway, it is very probable that the problem comes from there, and that the Thermo-Calc subroutine for getting diffusion data is not giving a correct warning. Thus, my advice is to switch off diffusion in graphite ("diagonal n") and to try again.

Bernd

[SuperX]

Dear Bernd:

Your advice really works! I have been troubled by this probelm for few days and I just cannot find the reason.
Thank you for your help!

Regards

[SuperX]

Dear Bernd:

I have got some problem in the "phase interaction date" section.I want to simulate the eutectic solidification of ductilt cast iron, but when it comes to the "Kinetic coefficient mu between phases LIQUID and 2" and "Interfacial energy between phases 0 and 2" , I do not know how to choose the number for graphit and for austenit.
It seems difficult to get these numbers from Google. And I have tried the input different numbers, and the results are quite different. It seems that these coefficients do have a great influence on the simulation results.Can you tell me how to set these coefficients correctly? Thanks a lot .

[Bernd]

Dear SuperX,

while the interface mobility is a mostly numerical parameter, the interface energy is physical (but probably unknown).

That means that the first one needs to be calibrated for diffusion-limited growth (see here). However, you should first remove the limits on the automatic time-stepping which cuts mobility and would confuse. You can use them later for performance optimisation once the numerical parameters are set up correctly.

With interface energy, a reasonable approach is to make an estimate (e.g. like you did), to compare with experimental results, and to adapt if it seems adequate.

Furthermore, you should remove the concentration limits which (as defined now) in best case have no effect, in worst case could be harmful.

Bernd

[SuperX]

Dear Bernd:
I used the method you told me. It works! And I have got some suitable parameter settings. Thankyou very much for your advise. Now I still have some questions that I can not solve myself, so I turn to you again for help.

1. The first one is about the appearance of grains. As far as I know the appearance of graphite in ductile cast iron should be a sphere. But my result is stars with 4 angles. I do not know which parameters control the appearance of grains.

2.The second one is about the separation of carbon element during the cooling process. According to the phase diagramm the maximal solubility of carbon in austenit decreases when the temperature drops. So some carbon will separate from austenit in form of graphite. How should this process be described? Should I set a new nucleation for this separation?

3. The third one is about TQ volumen control. I can not understand the function of "TQ volumen control". Also the molar volume of phases is required to be set, I think this parameters could influence the shrink, right? Are these parameters fixed nummers or should I try many times to find a suitable number?

Sorry I may have too many questions...but I still hope you can help me. Thanks again.
SuperX

[Bernd]

Hi SuperX,

Essentially, growth of spherical graphite in FeCSi is a quite complex process which involves a complex type of anisotropy, plastic deformation of the austenite shell once it encloses the graphite, not negligible volume changes, and fast interstitial diffusion of carbon through the austenite in direction of the growing graphite. Most of these effects cannot be directly treated using the standard MICRESS software. However, Janin is our expert on that topic in our development team, and she is working on exactly these problems.

Nevertheless, I try to give answers to your questions:

1.) If you see fourfold structures without explicitly applying such anisotropy, it most likely is due to the anisotropy of the numerical grid. To avoid or reduce this effect, everything is helpful which stabilizes the interface profile like increased resolution, interface stabilisation (optional parameter in the same line as interface energy), averaging of the driving force and correct adjustment of interface mobility. There is the special dG-option "smooth" which also may be helpful to reduce grid anisotropy.

2.) At the present stage it seems that you did not yet include the phase interaction between graphite and austenite. In fact, graphite will keep growing after being engulfed and consume the excess carbon in austenite.

3.) I don't know exactly what you mean by "TQ volumen control". MICRESS in its standard version does not take volume changes into account, and the input of molar volumes in connection to thermodynamic data serves only for converting the molar driving forces from the database to a driving force per volume. The optional keyword "volume" (in a new line after having specified the location and name of the .ges5 file) only allows specifying a value for a global pressure which is not standard pressure.
Janin works on a MICRESS version which really handles volume changes, but this is not yet available in the standard version.

Bernd

[SuperX]

Hello Bernd:

Thankyou very much for your help. According to your saying " Most of these effects cannot be directly treated using the standard MICRESS software", does this mean that my project(solidification of ductile cast iron) cannot be solved with Micress? That could not be worse to me if that is true....

SuperX

[janin]

Dear SuperX,
the major problem of simulating ductile cast iron is that growth of spheroidal graphite is controlled by interstitial diffusion of carbon through the austenite shell, while the whole shell including Fe as well as the solute elements are pushed away by the volumetric expansion of grahite.
If you neglect these mechanisms, you will find that the slow diffusing substitutional elements will control the graphite growth which is higly unrealistic, i.e. much too slow. Pragmatic solutions would be to simulate only the binary Fe-C_equivalent system or to manipulate the substitutional diffusion coefficients.

Since both solutions are not really convincing, we have develped a new volumetric expansion solver which will be part of the coming release.
However, you can already test the new model with your present Micress version 6.300.
I attach a simple example input file which illustrates how to simulate the growth of spheroidal graphite with an austenitic shell.

CastIronNodule_version6300_dri.txt

(28.15 KiB) Downloaded 328 times

You can also find a first presentation of this new model and simulation results in the web:
https://www.researchgate.net/publicatio ... egregation

(or ask me by email to send you a copy).

Best regards,
Janin

[SuperX]

Dear Janin:
You save me from my bachelor thesis! Really thanks~

Thanks Bernd again.

Best Wishes

Super X