Cementite dissolution problem

solid-solid phase transformations, influence of stresses and strains
Post Reply
Farnaz
Posts: 3
Joined: Thu Sep 03, 2020 1:34 pm
anti_bot: 333

Cementite dissolution problem

Post by Farnaz » Mon Sep 28, 2020 11:22 am

Hello,

Thanks for such an active forum.

I simulated austenitization from a ferrite+cementite matrix. Please find attached the initial microstructure and the simulation results.

In the initial microstructure, you see ferrite with distributed cementite. In the resulting picture, you see undissolved cementite in the austenite matrix.

But why is cementite blue (boundary?!?)? Does it affect the cementite fraction calculation in the software? Does it affect cementite dissolution in every iteration? How can I track the problem source?

Looking forward to hearing from you.

Best regards,
Farnaz
Attachments
Result.JPG
Result.JPG (109.75 KiB) Viewed 1772 times
Initial_microstructure.JPG
Initial_microstructure.JPG (134.88 KiB) Viewed 1772 times

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cementite dissolution problem

Post by Bernd » Mon Sep 28, 2020 12:16 pm

Hi Farnaz,

Indeed, the blue color (-1 value) of the .phas output indicates that the cementite has got "diffuse", i.e. is has no full phase fraction anymore. You should use the .frac* output in order to see the phase fraction distribution for each phase directly. You can activate this output e.g. by using

out_fraction 1 2 3

Further, you can look at .intf or .driv to monitor the interface and its thickness. Generally, numerical broadening of the interface leads to the effect of having such diffuse phases, which is clearly a numerical artifact. Generally, this happens when a strong gradient of the driving force exists over the interface thickness, which destabilizes the interface. The effect is generally facilitated by low grid resolution, high interface mobility and low interfacial energy, and can be avoided or reduced by using interface stabilisation (defining a stabilisation energy in the same line with the interface energy which can be up to 10 times bigger than the latter).

However, in your specific case, I would guess the problem comes from the slow-diffusing substitutional elements which are typically present in steels, while the cementite dissolution is driven by the fast-diffusing carbon. Especially with high amounts of substitutional elements like Cr or Mn, this easily leads to numerically unstable interfaces when "normal" redistribution is assumed. Therefore (and also for having correct results!), you should use either nple or para/paraTQ" in order to avoid that (I personally think that nple is the more realistic choice).

I would advise you to first try to switch on the nple model for the substitutional elements, assuming diffusion limited growth for the element C only. You can do that (if you do not already do) by using "redistribution_control", and using

nple

for all substitutional elements, and

normal mob_corr

for carbon, together with a sufficiently high interface mobility. You can apply this to fcc/bcc, fcc/cem and bcc/cem interfaces. Be careful to use nple only for elements which have a broad solubility range in all phases (and not e.g. for Si).

Bernd

Farnaz
Posts: 3
Joined: Thu Sep 03, 2020 1:34 pm
anti_bot: 333

Re: Cementite dissolution problem

Post by Farnaz » Fri Nov 27, 2020 5:17 pm

Dear Bernd,
Thanks for your comprehensive reply.

This material consist of 0.16 wt% C, 0.56 wt% Mn, 0.34 wt% Si, 3.3 wt% Ni, 1.4 wt% Cr ,and 0.17 wt% Mo. I used the “redistribution_control” option with “normal mob_corr” for carbon and “nple” for Cr (and/or Mn).
I received this error:

Start Composition and Limits for quasi-equilibrium
--------------------------------------------------
FE in FCC_A1: 39.0622 at% (>0 - 100.000at% )
C in FCC_A1: 0.910202 at% (>0 - 50.0000at% )
NI in FCC_A1: 58.0654 at% (>0 - 100.000at% )
CR in FCC_A1: 0.330584 at% (>0 - 100.000at% )
MN in FCC_A1: 0.458552 at% (>0 - 100.000at% )
MO in FCC_A1: 0.522536 at% (>0 - 100.000at% )
SI in FCC_A1: 0.650503 at% (>0 - 100.000at% )
FE in BCC_A2: 31.2375 at% (>0 - 100.000at% )
C in BCC_A2: 2.76584 at% (>0 - 75.0000at% )
NI in BCC_A2: 0.387728 at% (>0 - 100.000at% )
CR in BCC_A2: 28.8550 at% (>0 - 100.000at% )
MN in BCC_A2: 0.339596 at% (>0 - 100.000at% )
MO in BCC_A2: 36.0026 at% (>0 - 100.000at% )
SI in BCC_A2: 0.411794 at% (>0 - 100.000at% )
FE in M23C6: 40.9662 at% (>0 - 79.3103at% )
C in M23C6: 20.6897 at% ( 20.6897 - 20.6897at% )
NI in M23C6: 0.573234 at% (>0 - 79.3103at% )
CR in M23C6: 33.2551 at% (>0 - 79.3103at% )
MN in M23C6: 0.491690 at% (>0 - 79.3103at% )
MO in M23C6: 4.02403 at% (>0 - 10.3448at% )
SI in M23C6: 0.00000 at% (0 - 0at% )
FE in M7C3: 0.811056 at% (>0 - 70.0000at% )
C in M7C3: 30.0000 at% ( 30.0000 - 30.0000at% )
NI in M7C3: 1.29476 at% (>0 - 70.0000at% )
CR in M7C3: 64.3324 at% (>0 - 70.0000at% )
MN in M7C3: 0.972290 at% (>0 - 70.0000at% )
MO in M7C3: 1.13352 at% (>0 - 70.0000at% )
SI in M7C3: 1.45599 at% (>0 - 70.0000at% )


--> Force automatic start values

--> Force automatic start values
Initial concentration in the phases:
------------------------------------
FCC_A1, FE: 100.0000 wt%
FCC_A1, C: ************* wt%
FCC_A1, NI: ************* wt%
FCC_A1, CR: ************* wt%
FCC_A1, MN: ************* wt%
FCC_A1, MO: ************* wt%
FCC_A1, SI: ************* wt%
BCC_A2, FE: 95.30425 wt%
BCC_A2, C: 7.4620700E-04 wt%
BCC_A2, NI: 3.345000 wt%
BCC_A2, CR: 0.4950000 wt%
BCC_A2, MN: 0.4750000 wt%
BCC_A2, MO: 2.4000000E-02 wt%
BCC_A2, SI: 0.3560000 wt%
M23C6, FE: 47.31133 wt%
M23C6, C: 5.099434 wt%
M23C6, NI: 1.912851 wt%
M23C6, CR: 29.60411 wt%
M23C6, MN: 1.396531 wt%
M23C6, MO: 14.67574 wt%
M23C6, SI: 0.000000 wt%
M7C3, FE: 33.06049 wt%
M7C3, C: 8.675874 wt%
M7C3, NI: 0.2302791 wt%
M7C3, CR: 49.59574 wt%
M7C3, MN: 6.136814 wt%
M7C3, MO: 2.300803 wt%
M7C3, SI: 1.9790688E-10 wt%

Routine init calls routine initConcFeld
Routine init calls routine initTwidth

tWidth_max( FCC_A1: FCC_A1) = 4.9026786E-05 s
tWidth_max( BCC_A2: BCC_A2) = 2.4513393E-04 s
tWidth_max( BCC_A2: M23C6) = 0.1961071 s
tWidth_max( BCC_A2: M7C3) = 0.1961071 s
Maximal value for tWidth = 4.9026786E-05 s for phase-field solver
Updating of diffusion data from database...
Maximal value for tWidth = 7.6348767E-07 s for conc-field solver
Automatic time stepping (phase-field solver): decreased value for tWidth = 7.25313E-07 s
Initial value for tWidth = 7.25313E-07 s for automatic time stepping (phase-field solver)
Critical grain radius:
of phase M23C6 in phase BCC_A2 = 0.11952 / dT_unt [micrometers]
of phase M7C3 in phase BCC_A2 = 9.65084E-02 / dT_unt [micrometers]

End of routine init


Remaining license time: permanent

==================================================

Time t = 0.0000000 s
CPU-time: 2 s
Current phase-field solver time step = 7.25E-07 s
Average conc. of comp. C = 0.1615927 wt%
Average conc. of comp. NI = 3.2904707 wt%
Average conc. of comp. CR = 1.4074094 wt%
Average conc. of comp. MN = 0.5605272 wt%
Average conc. of comp. MO = 0.1738306 wt%
Average conc. of comp. SI = 0.3482130 wt%
Temperature at the bottom = 1000.0 K
Temperature gradient = 0.00000 K/cm
Fraction of phase LIQUID: 0.00000
Fraction of phase FCC_A1: 0.00000
Fraction of phase BCC_A2: 0.97242
Fraction of phase M23C6: 0.00919
Fraction of phase M7C3: 0.01839

Updating of diffusion data from database...

Intermediate output for t = 1.00000E-02 s
CPU-time: 135 s
Current phase-field solver time step = 1.67E-06 s
Average conc. of comp. C = 0.1615927, Variation = -0.0000000 wt%
Average conc. of comp. NI = 3.2904707, Variation = +0.0000000 wt%
Average conc. of comp. CR = 1.4074094, Variation = +0.0000000 wt%
Average conc. of comp. MN = 0.5605272, Variation = -0.0000000 wt%
Average conc. of comp. MO = 0.1738306, Variation = +0.0000000 wt%
Average conc. of comp. SI = 0.3482130, Variation = -0.0000000 wt%
Temperature at the bottom = 1000.1 K
Temperature gradient = 0.00000 K/cm

--> Force automatic start values
7
CALFUN:Error1 7
7
CALFUN Error1 not resolved!
7
CALFUN:Error1 7
Would you please help me to understand what is happening and what does this error means? Please let me know if I should provide more information.

Thanks in advance.

Best regards,
Farnaz

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Cementite dissolution problem

Post by Bernd » Fri Nov 27, 2020 7:07 pm

Hi Farnaz,

The error numbers are internal codes which tell us developers at which place in the code the problem occurs. In this case it seems that you are using the "paraTQ" option for one of the interfaces which exist in your simulation (BCC_A1/M23C6, BCC_A1/M7C3). Please note that you can use the options "nple", "para" and "paraTQ" only for elements which have a broad solubility range in both phases, so that partitionless transformation or overrunning of segregation peaks is permitted. In steels, this is typically the case only for interactions between fcc, bcc and cementite.

I would guess that this is the problem. For more details I would need to see your whole input file.

Best wishes

Bernd

Post Reply