Problem in coupling ThermoCalc with MICRESS

Discussion of numerical issues which are not directly linked to specific applications
Post Reply
kmukherj
Posts: 17
Joined: Wed Aug 12, 2009 3:22 pm
anti_bot: 333

Problem in coupling ThermoCalc with MICRESS

Post by kmukherj » Tue Mar 09, 2010 12:28 pm

Dear Dr. Böttger,

I tried to run a MICRESS code for an alloyed steel. I used ThermoCalc coupling. For creating the .GES5 file I used MOB2.

I am modelling FCC_A1 to BCC_A2 phase transformation at 540 K. But the BCC_A2 nuclei dissolve. After that I plotted the equilibrium phase diagram in ThermoCalc to see why the BCC_A2 nuclei are dissolving. Please find it attached with. I see there is a FCC_A1#2 line. I do not know what it is and whether it is responsible for the dissolution of the BCC_A2 nuclei.

Can you please help me in this regard?

Thanking you,

Yours sincerely,

Krishnendu Mukherjee
Attachments
Equilibrium phase diagram.JPG
Equilibrium phase diagram.JPG (51.27 KiB) Viewed 1687 times

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

Re: Problem in coupling ThermoCalc with MICRESS

Post by Bernd » Tue Mar 09, 2010 6:40 pm

Dear kmukhery,

what I understand is that you start the simulation at 540K with pure fcc phase. This is very ambitious, because at such low temperatures the diffusivities are extremely small, furthermore you may have extremely low equilibrium compositions. I never tried such a simulation and cannot know whether it is possible, but let's see!

In the phase diagram, there is also CEMENTITE and FCC_A1#2, which stands for MC carbides (In the database, FCC_A1 is used for describing as well fcc as MC!). How do you want to include them into the simulation? I guess, they are somehow part of the bainite structure...

Let us check what happens to the BCC nuclei. The first step would be to see whether the thermodynamic interaction between FCC and BCC is correctly initialized. If you start with pure fcc, the initialisation is done when you for the first time check for nucleation of bcc. Then, you get an output of the initial linearisation data in the .log file:

# The linearisation parameter of the phases FCC_A1#1/BCC_B2#1 are:
# --------------------------------------------------------------------------------
xxxx.xxxxx ! T0 [K]
xx.xxxx
...

Please search for this output in the .log file and paste it together with your next post in this forum thread! It allows to see whether under these extreme conditions the correct equilibrium is found!
Please check also the output on screen during nucleation: You should see the number of seeds which are created and the nucleation undercooling (which should be extremely high...).

Afterwards, we can see whether the problem is due to other circumstances like a wrong mobility or other numerical details.


Bernd

kmukherj
Posts: 17
Joined: Wed Aug 12, 2009 3:22 pm
anti_bot: 333

Re: Problem in coupling ThermoCalc with MICRESS

Post by kmukherj » Tue Mar 09, 2010 9:50 pm

Dear Dr. Böttger,

At this very low temperature lower Bainite forms. The carbides precipitate inside the Bainite. But at this point I do not have the idea to how to include them in the simulation.

Here is the portion of the .log file that you have requested for:


**********************************************
* Begining of simulation *
**********************************************

# The linearisation parameters of the phases FCC_A1/BCC_A2 are:
# -------------------------------------------------------------
533.00000 ! T0 [K]
-435.16159 ! dG [J/cm**3]
1.2348026 ! dSf+ [J/cm**3K]
-0.14033200 ! dSf- [J/cm**3K]
1383.6477 ! dH [J/cm3]
************* ! c0(C)/FCC_A1
2.65565686E-05 ! c0(C)/BCC_A2
************** ! c0(MN)/FCC_A1
2.91180215E-03 ! c0(MN)/BCC_A2
************** ! c0(SI)/FCC_A1
0.48886603 ! c0(SI)/BCC_A2
************* ! c0(CR)/FCC_A1
7.94948265E-02 ! c0(CR)/BCC_A2
-30.538420 ! m(C)/FCC_A1
11473108. ! m(C)/BCC_A2
-3.0205154 ! m(MN)/FCC_A1
5381.8774 ! m(MN)/BCC_A2
1.8889513 ! m(SI)/FCC_A1
18.940243 ! m(SI)/BCC_A2
1.4175082 ! m(CR)/FCC_A1
1114.0457 ! m(CR)/BCC_A2
-2.39220280E-02 ! dcdT(C)/FCC_A1
7.29774911E-07 ! dcdT(C)/BCC_A2
-6.08196342E-03 ! dcdT(MN)/FCC_A1
3.29172981E-05 ! dcdT(MN)/BCC_A2
4.15310729E-03 ! dcdT(SI)/FCC_A1
-2.22606803E-04 ! dcdT(SI)/BCC_A2
-5.28851599E-02 ! dcdT(CR)/FCC_A1
6.35265314E-04 ! dcdT(CR)/BCC_A2
# Minimum undercooling for stable growth, seed type 1: 5.183015 K [r=6.2499998E-02 mic.]

Seed number 1 set at time t = 1.00E-01 s
----------------------------------------
in user-defined area
Phase: 2 (BCC_A2)
Seed type: 1
Temperature at the bottom = 533.00 K
Undercooling = 352.42 K
Grain number = 2

Intermediate output for t = 0.10000 s
CPU-time: 0 s
Current phase-field solver time step = 1.00E+20 s

...................................................................................................................................



Thanking you,

Yours sincerely,

Krishnendu

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

Re: Problem in coupling ThermoCalc with MICRESS

Post by Bernd » Wed Mar 10, 2010 12:54 pm

Dear Krishnendu,

The linearisation data seem to be quite reasonable, i.e. there is a strong driving force dG for formation of BCC, and the values of the equilibrium compositions (at least those which you show) and the transformation entropy dSf+ and dSf- (to be more exact: the partial derivatives of dG to the temperature) have reasonable values.
But the system shows strong "demixing" tendencies: If a component shows opposite signs for the product dSf+ * m_FCC and dSf- * m_BCC, the multibinary extrapolation scheme, which is applied in MICRESS during redistribution of the element between the phases, is not working properly. In such a case it is necessary to define one of the two phases for those elements as stoichiometric. Then, even if the component is not really stoichiometric, redistribution of the elements is possible. One will typically take the phase with the higher value of the slope for the stoichiometric condition, because then the approximation as stoichiometric phase is not so drastic. In your case, Si and Cr (elements 3 and 4) should be declared as stoichiometric in BCC (phase 2):

# List of phases and components which are stoichiometric:
# phase and component(s) numbers
# List of concentration limits:
# <Limits>, phase number and component number
# End with 'no_more_stoichio' or 'no_stoichio'
2 3 4
no_more_stoichio

Please try that and send again the linearisation parameter (to be sure..). Principially, this could already solve your problem, but in a such difficult case like yours I expect more trouble...


Bernd

kmukherj
Posts: 17
Joined: Wed Aug 12, 2009 3:22 pm
anti_bot: 333

Re: Problem in coupling ThermoCalc with MICRESS

Post by kmukherj » Fri Mar 12, 2010 6:35 pm

Dear Dr. Böttger,

Unfortunately the BCC_A2 nucleus dissolved again.

Here is the linearisation parameter:


**********************************************
* Begining of simulation *
**********************************************

# The linearisation parameters of the phases FCC_A1/BCC_A2 are:
# -------------------------------------------------------------
533.00000 ! T0 [K]
-435.16159 ! dG [J/cm**3]
1.2348026 ! dSf+ [J/cm**3K]
0.20616648 ! dSf- [J/cm**3K]
1383.6477 ! dH [J/cm3]
************* ! c0(C)/FCC_A1
2.65565686E-05 ! c0(C)/BCC_A2
***************** ! c0(MN)/FCC_A1
2.91180215E-03 ! c0(MN)/BCC_A2
**************** ! c0(SI)/FCC_A1
0.48886603 ! c0(SI)/BCC_A2
************* ! c0(CR)/FCC_A1
7.94948265E-02 ! c0(CR)/BCC_A2
-30.538420 ! m(C)/FCC_A1
-6192803.0 ! m(C)/BCC_A2
-3.0205154 ! m(MN)/FCC_A1
-3242.0723 ! m(MN)/BCC_A2
1.8889513 ! m(SI)/FCC_A1
-999.99994 ! m(SI)/BCC_A2
1.4175082 ! m(CR)/FCC_A1
-999.99994 ! m(CR)/BCC_A2
-1.64906457E-02 ! dcdT(C)/FCC_A1
7.29774911E-07 ! dcdT(C)/BCC_A2
-5.02055790E-03 ! dcdT(MN)/FCC_A1
3.29172981E-05 ! dcdT(MN)/BCC_A2
1.91814377E-11 ! dcdT(SI)/FCC_A1
-2.22606803E-04 ! dcdT(SI)/BCC_A2
9.68477989E-12 ! dcdT(CR)/FCC_A1
6.35265314E-04 ! dcdT(CR)/BCC_A2
# Minimum undercooling for stable growth, seed type 1: 5.183015 K [r=6.2499998E-02 mic.]

Seed number 1 set at time t = 1.00E-01 s
----------------------------------------
in user-defined area
Phase: 2 (BCC_A2)
Seed type: 1
Temperature at the bottom = 533.00 K
Undercooling = 352.42 K
Grain number = 2


------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


Can you please explain dSf+ and dSf- in little more detail? And also can you please explain what is dcdT(C)/FCC_A1?

I think now we can shift our discussion to a published study of the same phase transformation. : Bainite formation kinetics in high carbon alloyed steel: N. V. Luzginova, L. Zhao, and J. Sietsma, Materials science and engineering A, 481-482, 2008, 766-769. I calculated the linearisation parameter for this published chemistry and found the same problem of dissolution of BCC_A2 nucleus. And then I can put all the numerical details in forum. Please let me know if you agree with this.


Thanking you,

Yours sincerely,

Krishnendu

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

Re: Problem in coupling ThermoCalc with MICRESS

Post by Bernd » Sun Mar 14, 2010 1:10 pm

Dear Krishnendu,

of course, if you wish, we can switch to another alloy and a linearized phase diagram description where you can show all the details in the forum. I wanted to check the thermodynamics first, because this is reasonable to be done first, but as we found out, thermodynamics are not the problem! The next things to check are details about nucleation, phase interaction data and further numerical details.

If you want to switch, please give us the linearisation parameter in your next post!

dSf+ and dSf- are the derivatives of dG to the temperature, keeping either the composition of the first or the second phase (fcc, bcc) fixed. A different sign in dSf+ and dSf- means that the solubilty of the two phases is reduced by lowering the temperature, like it is the case e.g. for the solid phases in a typical eutectic reaction.

dcdT are the derivatives of the quasi-equilibrium compostion of the two phases with temperature. They are used for a better interpolation between relinearisations.

Bernd

Post Reply