Heat Treatment simulation of Ni-alloy

solid-solid phase transformations, influence of stresses and strains
Bernd
Posts: 1221
Joined: Mon Jun 23, 2008 9:29 pm

Re: Heat Treatment simulation of Ni-alloy

Post by Bernd » Thu Mar 22, 2018 3:44 pm

Hi Chamara,

with "restart structure only" MICRESS has to reinitialize the interactions between all phases and calculate the linearisation parameters of the initial microstructure. The errors which occur are due to that process. You should use the keyword "input" for specification of initial concentrations, so that you can give reasonable start compositions for all phases. You also can adapt the temperature for initial equilibrium in order to make this process easier.
Please also check whether the initial microstructure looks like expected. The process of initialisation can take quite long depending on the size and complexity of the read microstructure and on the relinearisation mode which you use (local, global, ...).

Bernd

CharMIC
Posts: 213
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Heat Treatment simulation of Ni-alloy

Post by CharMIC » Thu Mar 22, 2018 3:54 pm

Hej Bernd,

Thanks for the input. But in the initial microstructure (i.e.End of solidification simulation microstructure) the element are segregated in dendrite core and within the interdendritic regions. So how can I input a composition for the segregated microstructure?

Do I just input some kind of average value?

BR
Bernd

Bernd
Posts: 1221
Joined: Mon Jun 23, 2008 9:29 pm

Re: Heat Treatment simulation of Ni-alloy

Post by Bernd » Thu Mar 22, 2018 9:37 pm

Dear Chamara,

the strategy is open to your own experiences - it is only about providing good start values for Thermo-Calc. In most cases the first trial will be to use average compositions, and think about other strategies only if the initial equilibrium fails...

Bernd

CharMIC
Posts: 213
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Heat Treatment simulation of Ni-alloy

Post by CharMIC » Fri Mar 23, 2018 1:46 pm

Hej Bernd,

So If I understood correctly, the input parameters are just to Thermo-calc to get started. But in the calculations it use the composition fields read by the initial microstructure. So the manually inputted initial values will not affect the results?

BR
Chamara

Bernd
Posts: 1221
Joined: Mon Jun 23, 2008 9:29 pm

Re: Heat Treatment simulation of Ni-alloy

Post by Bernd » Fri Mar 23, 2018 3:54 pm

Hej Chamara,

This is correct (at least if the complete initial microstructure consists of contents of .rest files).

Bernd

CharMIC
Posts: 213
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Heat Treatment simulation of Ni-alloy

Post by CharMIC » Mon Apr 23, 2018 8:56 am

Hej Bernd,

I am getting the following error in a heat treatment simulation after 60 sec. In the initial microstructure consist of Gamma phase and Laves phase. Heat treatment is done at 1280K.

In the .GES5 file major constitutes has been adjusted according to the suggestions that you have mentioned (http://board.micress.de/viewtopic.php?f=9&t=563). Also I am using stoichio condition for the all the elements in LAVES phase except Ni. And Limits criteria for the NB in LAVES phase.I am using multi llllll option in the gamma phase.

I have done the same simulation with multii gggggg option. There I don't get any error.

Error seems to come due to the violation of the limits criteria for element 5 (NB) in phase 2 (LAVES). But what I dont understand is why I am not getting the error when I use multi gggggg option. Also how can I avoid this problem?

BR
Chamara
Attachments
HeatTreatment_After60s.JPG
HeatTreatment_After60s.JPG (61.23 KiB) Viewed 3611 times

Bernd
Posts: 1221
Joined: Mon Jun 23, 2008 9:29 pm

Re: Heat Treatment simulation of Ni-alloy

Post by Bernd » Mon Apr 23, 2018 10:10 am

Hi Chamara,

The first step should be to check whether there is any sign of instability of the diffusion solver. This can be clearly diagnosed by the formation of a chess pattern in the concentration distribution. Due to the interrelations between compositions when transforming from atomic to weight percent, this pattern typically is visible in all conc* outputs if wt% is used as concentration unit. The chess pattern rapidly evolves so that in most cases INF or NAN values cover big parts of the domain after short time.

If this is the case, extrapolation of diffusion data by using the "l" option has obviously led to an unstable diffusion matrix, which can very easily happen if the phase composition is close to a miscibility gap. You can try to limit the extrapolation range by additionally defining for each diffusion term, e.g.

1 1 maxfactor_local 2.0

which may help. Some more information you can find here.

Bernd

CharMIC
Posts: 213
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Heat Treatment simulation of Ni-alloy

Post by CharMIC » Mon Apr 23, 2018 10:30 am

Many Thanks Bernd. Is this the chess pattern that you are referring.
Attachments
Chess Pattern.JPG
Chess Pattern.JPG (54.19 KiB) Viewed 3608 times

Bernd
Posts: 1221
Joined: Mon Jun 23, 2008 9:29 pm

Re: Heat Treatment simulation of Ni-alloy

Post by Bernd » Mon Apr 23, 2018 10:41 am

Difficult to say...

Maybe, you should also check the .c*pha1 output (i.e. the phase composition inside fcc) which is in at% and which shows you also the fcc compositions inside the interfaces. If you see similar patterns there also, it could be a very localized instability of diffusion just "behind" the laves particles...

Bernd

CharMIC
Posts: 213
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Heat Treatment simulation of Ni-alloy

Post by CharMIC » Tue Apr 24, 2018 11:17 am

Hej Bernd,

Your suggestion worked :) (1 1 maxfactor_local 2.0)

BR
Chamara

Post Reply