Dear Bernd,
Could you explain a bit why there are 3 different seed types in the CMSX4 example to simulate the precipitation of gamma' phase. I am a bit confuse.
BR
Chamara
Data for further nucleation
Re: Data for further nucleation
Hi Chamara,
first you should clarify which version number you mean - we are just sending out MICRESS6-4, and the CMSX4-example has been modified
Anyway, the 3 seed types are still there:
1.) Nucleation of gamma' at the liquid/gamma interface
2.) Nucleation of gamma at the liquid/gamma' interface (to enable eutectic growth or to help solidification of liquid pockets enclosed inside gamma' phase
3.) Transformation of rest liquid to gamma below 1400°C in order to obtain a completely solid microstructure (this is important e.g. if one wants to simulate heat treatment afterwards).
Best wishes
Bernd
first you should clarify which version number you mean - we are just sending out MICRESS6-4, and the CMSX4-example has been modified
Anyway, the 3 seed types are still there:
1.) Nucleation of gamma' at the liquid/gamma interface
2.) Nucleation of gamma at the liquid/gamma' interface (to enable eutectic growth or to help solidification of liquid pockets enclosed inside gamma' phase
3.) Transformation of rest liquid to gamma below 1400°C in order to obtain a completely solid microstructure (this is important e.g. if one wants to simulate heat treatment afterwards).
Best wishes
Bernd
Re: Data for further nucleation
Hej Bernd,
I was referring to the MICRESS 6.3
In the CMSX4 example, the γ' has not been modeled as a cubic symmetry phase. What is the reason behind this. ?
BR
Chamara
I was referring to the MICRESS 6.3
In the CMSX4 example, the γ' has not been modeled as a cubic symmetry phase. What is the reason behind this. ?
BR
Chamara
Re: Data for further nucleation
Hi Chamara,
The reason is just that it does not make a difference here. A further reason is, if we set gamma' to isotropic, the seeds have no different orientations, and we can more easily "categorize" them for improved performance.
However, there should be no problem with setting gamma' to anisotropic!
Bernd
The reason is just that it does not make a difference here. A further reason is, if we set gamma' to isotropic, the seeds have no different orientations, and we can more easily "categorize" them for improved performance.
However, there should be no problem with setting gamma' to anisotropic!
Bernd
Re: Data for further nucleation
Hej Bernd,
I have a question regarding define neucliation parameter for precipitation of carbides in Nickel alloy.
Usually carbides will precipitate before formation of gamma matrix and with the movement of the liquid/gamma interface, the formed carbides will push towards the inter dendritic region.
If I defined the liquid/gamma interface as the nucleation site for the carbides, will it be moved with the advancing interface rather than trapped within the matrix?
BR
Chamara
I have a question regarding define neucliation parameter for precipitation of carbides in Nickel alloy.
Usually carbides will precipitate before formation of gamma matrix and with the movement of the liquid/gamma interface, the formed carbides will push towards the inter dendritic region.
If I defined the liquid/gamma interface as the nucleation site for the carbides, will it be moved with the advancing interface rather than trapped within the matrix?
BR
Chamara
Re: Data for further nucleation
Dear Chamara,
in Ni-base alloys there are two kinds of secondary carbides which can be distinguished: They are either eutectic (script like structures) or blocky. If they are "eutectic", they probably have nucleated at the solid-liquid interface, and there was a cooperative coupled growth. Otherwise they nucleated inside the bulk of the liquid phase, so that a divorced coupled growth occured.
Exactly that you should do in MICRESS: If you want to have cooperative eutectic growth, nucleate at interfaces, if you want blocky carbides, you nucleate in the bulk. The best way to do the latter is to use the seed-density model with an assumed density distribution.
Bernd
in Ni-base alloys there are two kinds of secondary carbides which can be distinguished: They are either eutectic (script like structures) or blocky. If they are "eutectic", they probably have nucleated at the solid-liquid interface, and there was a cooperative coupled growth. Otherwise they nucleated inside the bulk of the liquid phase, so that a divorced coupled growth occured.
Exactly that you should do in MICRESS: If you want to have cooperative eutectic growth, nucleate at interfaces, if you want blocky carbides, you nucleate in the bulk. The best way to do the latter is to use the seed-density model with an assumed density distribution.
Bernd
Re: Data for further nucleation
Hej Bernd,
Many Thanks.
Are there any tips when I defined seed density profile? like eg: Relative to the revolution and cooling conditions?
Some times I do not get any nucleation when I use the seed density model.
Chamara
Many Thanks.
Are there any tips when I defined seed density profile?
like eg: Relative to the revolution and cooling conditions?Some times I do not get any nucleation when I use the seed density model.
Chamara
Re: Data for further nucleation
Dear Chamara,
The seed-density distribution is a property of the melt and depends on its production history. In so far, there is no way of guessing, if no experimental data is available.
In a strict way, it would be possible to determine the seed-density distribution if sufficient trustful experimental information (e.g. grain sizes for many different cooling rates) is available. However, this typically is not the case, and even then it would be a quite difficult process.
There is, however, another way of using the seed_density model in a qualitative manor: One just invents a distribution (or even a single seed class) and changes it until it fits in what he/she expects to get. In this case, the output of the actual numbers of seed points per class in the .log file is an important help. Even when used in this way, typically more realistic structures can be obtained than when using the normal "bulk" nucleation.
Bernd
The seed-density distribution is a property of the melt and depends on its production history. In so far, there is no way of guessing, if no experimental data is available.
In a strict way, it would be possible to determine the seed-density distribution if sufficient trustful experimental information (e.g. grain sizes for many different cooling rates) is available. However, this typically is not the case, and even then it would be a quite difficult process.
There is, however, another way of using the seed_density model in a qualitative manor: One just invents a distribution (or even a single seed class) and changes it until it fits in what he/she expects to get. In this case, the output of the actual numbers of seed points per class in the .log file is an important help. Even when used in this way, typically more realistic structures can be obtained than when using the normal "bulk" nucleation.
Bernd
Re: Data for further nucleation
Hej Bernd,
Thanks as always for detail explanations. I have another question regarding simulating the precipitation of gamma'' and transforming them to delta phase at long exposure times at elevated temperatures..
Can this be handled using MICRESS. Gamma'' is rather quite small in size when compare to the grid resolution. So Any advice on how to set up the nucleation data to precipitate gamma'' and then transformation of them to delta at long exposure times.
BR
Chamara
Thanks as always for detail explanations. I have another question regarding simulating the precipitation of gamma'' and transforming them to delta phase at long exposure times at elevated temperatures..
Can this be handled using MICRESS. Gamma'' is rather quite small in size when compare to the grid resolution. So Any advice on how to set up the nucleation data to precipitate gamma'' and then transformation of them to delta at long exposure times.
BR
Chamara
Re: Data for further nucleation
Hi Chamara,
This is a general problem: If the effects you want to simulate depend strongly on the microstructure of the very fine particles, then there is no other possibility than to use very high resolution, and thus simulate only a very small region. Otherwise, it could be possible to take some approximation and work at lower resolution.
I have just summed up my recent thoughts about the latter case there. For further discussion please answer there, because here we are in the "Solidification" board.
Bernd
This is a general problem: If the effects you want to simulate depend strongly on the microstructure of the very fine particles, then there is no other possibility than to use very high resolution, and thus simulate only a very small region. Otherwise, it could be possible to take some approximation and work at lower resolution.
I have just summed up my recent thoughts about the latter case there. For further discussion please answer there, because here we are in the "Solidification" board.
Bernd