Hi Parimal,
to your first question: The second value after the interfacial energy is an optional parameter for interface stabilisation. Simply speaking, the interface energy in our phase-field model at the same time stabilizes the interface profile, i.e. prevents that it gets destroyed by strong gradients of driving force. Interface stabilisation provides a possibility to increase interface stability without increasing interface energy. As such, it is only a numerical parameter which can be chosen up to typically 10 times higher as the real interfacial energy.
The interfacial energy, if unknown, can be estimated using Thermo-Calc. Honestly, I am not able to judge how good these values are.
second question: If you assume that at the high cooling rates of the process you don't have nucleation of gamma', you do not need the first two nucleation types (the first is nucleation of gamma', the second was to ensure that liquid which was enclosed by gamma'could still solidify.
The third nucleation type will still be important: It switches rest liquid (which does not easily vanish for numerical reasons) to fcc. Something similar (maybe at different temperature) you may still need.
However, be careful with omitting gamma': This may be unrealistic and lead to numerical issues when the fcc phase accidentally switches to gamma' (a similar problem as with composition sets). I would expect that - if secondary gamma' really does not form due to the high cooling rate - it won't form with nucleation switched on. Switching it off means flying blind...
Bernd
CMSX4
-
parimalmaity
- Posts: 22
- Joined: Tue Mar 07, 2017 10:50 am
- anti_bot: 333
Re: CMSX4
Hello Bernd,
Thank you for your reply. I am sorry my intension is not to remove any phases. The reason I asked was to know if I only simulate one phase and slowly adding other phases into the system.
Could you please answer the question no 3 in last post?
Regards
Parimal
Thank you for your reply. I am sorry my intension is not to remove any phases. The reason I asked was to know if I only simulate one phase and slowly adding other phases into the system.
Could you please answer the question no 3 in last post?
Regards
Parimal
Re: CMSX4
Dear Parimal,
That is exactly the point I wanted to make: One may think ,to make it easier for the beginning, not take into account some phases. However, in some cases, this makes things more complicated, if the system then moves to concentration regions which are strongly metastable. A typical error e.g. is to simulate solidification of a superalloy containing carbon and - in a first instance - leave out carbide phases. What happens is that carbon is accumulating in the interdendritic rest liquid leading eventually to huge numerical problems.
A similar effect could happen when omitting secondary gamma' in typical SX superalloys: The driving force for gamma' formation could exceed that of gamma, and the phase could switch to the wrong description. The same typically happens also in Thermo-Calc if one of the 2 descriptions (ordered or disordered) are explicitly forbidden.
Sorry that I overlooked your third question:
The "seed_undercooling" nucleation model is very phenomenological, and the parameters like shield distance and time, nucleation distance, frequency of checking do not have a direct physical meaning but allow the user to design nucleation such as to resemble the experimentally observed behaviour. The shield distance, for example, can be correlated with the average distance of particles seen in the microstructure. Checking time and shield time should be in accordance with cooling rate to allow for a realistic nucleation sequence.
The only physical parameter here is the critical undercooling, but this is typically unknown. In principle, it can be derived from experiments, or from nucleation models which deliver a critical radius. However, such models are not available for nucleation at interfaces. Thus, it finally is left to the user to estimate those values. Fortunately, the critical undercooling in most cases has no strong influence on the results, because the solutal undercooling for secondary phases increases strongly once the temperature is falling below its liquidus temperature. In general, the critical undercooling should not be chosen too small, otherwise seeds may not be able to grow if undercooling is not sufficient to overcome curvature.
The input of the temperature range for nucleation in most cases is just used for optimization purposes, i.e. not to lose time with checking for nucleation at temperatures where you are sure that there will be no nucleation.
Bernd
That is exactly the point I wanted to make: One may think ,to make it easier for the beginning, not take into account some phases. However, in some cases, this makes things more complicated, if the system then moves to concentration regions which are strongly metastable. A typical error e.g. is to simulate solidification of a superalloy containing carbon and - in a first instance - leave out carbide phases. What happens is that carbon is accumulating in the interdendritic rest liquid leading eventually to huge numerical problems.
A similar effect could happen when omitting secondary gamma' in typical SX superalloys: The driving force for gamma' formation could exceed that of gamma, and the phase could switch to the wrong description. The same typically happens also in Thermo-Calc if one of the 2 descriptions (ordered or disordered) are explicitly forbidden.
Sorry that I overlooked your third question:
The "seed_undercooling" nucleation model is very phenomenological, and the parameters like shield distance and time, nucleation distance, frequency of checking do not have a direct physical meaning but allow the user to design nucleation such as to resemble the experimentally observed behaviour. The shield distance, for example, can be correlated with the average distance of particles seen in the microstructure. Checking time and shield time should be in accordance with cooling rate to allow for a realistic nucleation sequence.
The only physical parameter here is the critical undercooling, but this is typically unknown. In principle, it can be derived from experiments, or from nucleation models which deliver a critical radius. However, such models are not available for nucleation at interfaces. Thus, it finally is left to the user to estimate those values. Fortunately, the critical undercooling in most cases has no strong influence on the results, because the solutal undercooling for secondary phases increases strongly once the temperature is falling below its liquidus temperature. In general, the critical undercooling should not be chosen too small, otherwise seeds may not be able to grow if undercooling is not sufficient to overcome curvature.
The input of the temperature range for nucleation in most cases is just used for optimization purposes, i.e. not to lose time with checking for nucleation at temperatures where you are sure that there will be no nucleation.
Bernd
-
parimalmaity
- Posts: 22
- Joined: Tue Mar 07, 2017 10:50 am
- anti_bot: 333
Re: CMSX4
Hello Bernd,
Thank you.
There is one more observation in seed type #3
# Input for seed type 3:
# ----------------------
# Type of 'position' of the seeds?
# Options: bulk region interface triple quadruple [restrictive]
interface
# Phase of new grains (integer) [unresolved|add_to_grain]?
1 add_to_grain
# Which grain number the new grains shall be added to?
# # Options: grain_number (number) parent_grain new_set
parent_grain
# Reference phase (integer) [min. and max. fraction (real)]?
0
# Substrate phase [2nd phase in interface]?
# (set to 0 to disable the effect of substrate curvature)
0
# maximum number of new nuclei 3?
100000
# Grain radius [micrometers]?
0.00000
# Choice of growth mode:
# Options: stabilisation analytical_curvature
stabilisation
# min. undercooling [K] (>0)?
-1.00000E+05
# Determination of nuclei orientations?
# Options: random randomZ fix range parent_relation
random
# Shield effect:
# Shield time [s] [shield phase or group number] ?
0.0000
# Shield distance [micrometers] [ nucleation distance [micrometers] ]?
1.000
# Shall categorization be applied to this seed type?
# Options: categorize {Number} no_categorize
no_categorize
# Nucleation range
# min. nucleation temperature for seed type 3 [K]
0.000000
# max. nucleation temperature for seed type 3 [K]
1400.000
# Time between checks for nucleation? [s]
# Options: constant from_file
constant
# Time interval [s]
1.0000
# Shall random noise be applied?
# Options: nucleation_noise no_nucleation_noise
no_nucleation_noise
Question: The minimum undercooling is -1.00000E+05. Could you please explain why so high undercooling?
regards
Parimal
Thank you.
There is one more observation in seed type #3
# Input for seed type 3:
# ----------------------
# Type of 'position' of the seeds?
# Options: bulk region interface triple quadruple [restrictive]
interface
# Phase of new grains (integer) [unresolved|add_to_grain]?
1 add_to_grain
# Which grain number the new grains shall be added to?
# # Options: grain_number (number) parent_grain new_set
parent_grain
# Reference phase (integer) [min. and max. fraction (real)]?
0
# Substrate phase [2nd phase in interface]?
# (set to 0 to disable the effect of substrate curvature)
0
# maximum number of new nuclei 3?
100000
# Grain radius [micrometers]?
0.00000
# Choice of growth mode:
# Options: stabilisation analytical_curvature
stabilisation
# min. undercooling [K] (>0)?
-1.00000E+05
# Determination of nuclei orientations?
# Options: random randomZ fix range parent_relation
random
# Shield effect:
# Shield time [s] [shield phase or group number] ?
0.0000
# Shield distance [micrometers] [ nucleation distance [micrometers] ]?
1.000
# Shall categorization be applied to this seed type?
# Options: categorize {Number} no_categorize
no_categorize
# Nucleation range
# min. nucleation temperature for seed type 3 [K]
0.000000
# max. nucleation temperature for seed type 3 [K]
1400.000
# Time between checks for nucleation? [s]
# Options: constant from_file
constant
# Time interval [s]
1.0000
# Shall random noise be applied?
# Options: nucleation_noise no_nucleation_noise
no_nucleation_noise
Question: The minimum undercooling is -1.00000E+05. Could you please explain why so high undercooling?
regards
Parimal
Re: CMSX4
Hi Parimal,
You remember I told you that the third seed type was to switch small rests of liquid to fcc. Under these conditions, I do not respect the nucleation undercooling, I just want to switch the phase once the maximum temperature which I defined is reached. In order that a possibly missing driving force does not interfere, I define a large negative critical undercooling.
The option "add_to_grain" generally nucleates grains which are already existing. This gives me the opportunity either to put an instance of the same grain to another place and/or to redefine the properties of a grain. As in our case the same grain already exists at this place (grain 0), nothing happens apart from the property redefinition (which was my goal here).
Bernd
You remember I told you that the third seed type was to switch small rests of liquid to fcc. Under these conditions, I do not respect the nucleation undercooling, I just want to switch the phase once the maximum temperature which I defined is reached. In order that a possibly missing driving force does not interfere, I define a large negative critical undercooling.
The option "add_to_grain" generally nucleates grains which are already existing. This gives me the opportunity either to put an instance of the same grain to another place and/or to redefine the properties of a grain. As in our case the same grain already exists at this place (grain 0), nothing happens apart from the property redefinition (which was my goal here).
Bernd
-
parimalmaity
- Posts: 22
- Joined: Tue Mar 07, 2017 10:50 am
- anti_bot: 333
Re: CMSX4
Hello Bernd,
Thank you for your reply.
I have few more doubts and need your help in understanding.
Question 1:
When I look at thermo-calc equilibrium results, first from liquid it is forming gamma and then from gamma it is converted to gamma prime, gamma double prime and so on...But, in CMSX4 simulation nucleation setup I can see that gamma prime is forming from liquid and interface of gamma. When gamma prime will appear there should not be any liquid in the domain. I guess ideally gamma prime or double prime should appear from bulk gamma. Could you please explain? In principle, it should go from Liquid......>Gamma and Gamma.....>Gamma prime, double prime and so on.....
Question 2:
If I would like to do the similar setup what I need to do in nucleation?
Question 3:
In the simulation, I have observed that liquid temperature drops below solidus temperature but still liquid remains in domain. Is it due to directional solidification simulation? If I would like to perform in reality what I need to do? Could you please help in understanding?
Regards
Parimal
Thank you for your reply.
I have few more doubts and need your help in understanding.
Question 1:
When I look at thermo-calc equilibrium results, first from liquid it is forming gamma and then from gamma it is converted to gamma prime, gamma double prime and so on...But, in CMSX4 simulation nucleation setup I can see that gamma prime is forming from liquid and interface of gamma. When gamma prime will appear there should not be any liquid in the domain. I guess ideally gamma prime or double prime should appear from bulk gamma. Could you please explain? In principle, it should go from Liquid......>Gamma and Gamma.....>Gamma prime, double prime and so on.....
Question 2:
If I would like to do the similar setup what I need to do in nucleation?
Question 3:
In the simulation, I have observed that liquid temperature drops below solidus temperature but still liquid remains in domain. Is it due to directional solidification simulation? If I would like to perform in reality what I need to do? Could you please help in understanding?
Regards
Parimal
Re: CMSX4
Dear Parimal,
Q1: Thermo-Calc equilibrium is a very bad model for solidification. If you do a Scheil simulation with Thermo-Calc (which is a much better approximation), you will see that gamma' is formed before end of solidification. This you would also see in experiments (as cast microstructure) where the coarse gamma' eutectic indicates that it has formed from the liquid phase and not from solid.
Apart from that, there is gamma' nucleation in the solid also which appears later and which was not included in the CMSX4_dri example.
Q2: see Q1.
If you further want to include nucleation of gamma' in the solid at lower temperatures, you need an extra seed type for that. However, there is a fundamental problem with length-scale because these particles are typically far below 1µm in size. Therefore, you need a proper approximation (see e.g. B. Böttger, M. Apel, B.Laux, S. Piegert, "Detached Melt Nucleation during Diffusion Brazing of a Technical Ni-based Superalloy: A Phase-Field Study", 2015 IOP Conf. Ser.: Mater. Sci. Eng. 84 012031, http://dx.doi.org/10.1088/1757-899X/84/1/012031).
Q3: The same as Q1. The better comparison is Scheil, but even that may be bad beyond a fraction solid of 99%...
Bernd
Q1: Thermo-Calc equilibrium is a very bad model for solidification. If you do a Scheil simulation with Thermo-Calc (which is a much better approximation), you will see that gamma' is formed before end of solidification. This you would also see in experiments (as cast microstructure) where the coarse gamma' eutectic indicates that it has formed from the liquid phase and not from solid.
Apart from that, there is gamma' nucleation in the solid also which appears later and which was not included in the CMSX4_dri example.
Q2: see Q1.
If you further want to include nucleation of gamma' in the solid at lower temperatures, you need an extra seed type for that. However, there is a fundamental problem with length-scale because these particles are typically far below 1µm in size. Therefore, you need a proper approximation (see e.g. B. Böttger, M. Apel, B.Laux, S. Piegert, "Detached Melt Nucleation during Diffusion Brazing of a Technical Ni-based Superalloy: A Phase-Field Study", 2015 IOP Conf. Ser.: Mater. Sci. Eng. 84 012031, http://dx.doi.org/10.1088/1757-899X/84/1/012031).
Q3: The same as Q1. The better comparison is Scheil, but even that may be bad beyond a fraction solid of 99%...
Bernd
-
parimalmaity
- Posts: 22
- Joined: Tue Mar 07, 2017 10:50 am
- anti_bot: 333
Re: CMSX4
Hello Bernd,
Thank you for your reply and reference.
Only I am still not clear is that during solidification when liquid phase temperature drops down to solidus temperature and liquid still remains as liquid means something wrong setup in my simulation.
So, my question is what are check points if I can see temperature below solidus and still liquid remains in the system. I can understand there will be liquid in the inter-dendritic region where liquid is Ni-rich and hence will be solidified in less temperature than equilibrium (/initial) composition solidus temperature. But, above or far from the dendrite top region where composition remains almost unchanged (assumed) should be solidified below solidus temperature. If not, what condition I need to apply to make it realistic.
Regards
Parimal
Thank you for your reply and reference.
Only I am still not clear is that during solidification when liquid phase temperature drops down to solidus temperature and liquid still remains as liquid means something wrong setup in my simulation.
So, my question is what are check points if I can see temperature below solidus and still liquid remains in the system. I can understand there will be liquid in the inter-dendritic region where liquid is Ni-rich and hence will be solidified in less temperature than equilibrium (/initial) composition solidus temperature. But, above or far from the dendrite top region where composition remains almost unchanged (assumed) should be solidified below solidus temperature. If not, what condition I need to apply to make it realistic.
Regards
Parimal
Re: CMSX4
Dear Parimal,
when solidification not finishes although temperature is already low enough it may have various reasons. One of the most frequent is that some phases are missing which are "needed" for solidification. I explained that before at the example of the Carbides. A good check is to compare which phases appear in Thermo-Calc Scheil simulation.
But also the way how nucleation is checked can be the culprit. Sometimes, there are enclosed liquid pockets, and you also need all those phases inside (sometimes even the primary fcc phase may be missing somewhere).
Another reason may be that a phase is existing, but interface mobility is too low.
In many cases, the last rests (e.g. <0.01%) of liquid are very difficult to remove, especially if diffusion in the solid phases is slow. Then it may be a pragmatic solution just to "switch" this rest to fcc once a sufficiently low temperature is reached. This can be done using the "add_to_grain" option with nucleation.
Bernd
when solidification not finishes although temperature is already low enough it may have various reasons. One of the most frequent is that some phases are missing which are "needed" for solidification. I explained that before at the example of the Carbides. A good check is to compare which phases appear in Thermo-Calc Scheil simulation.
But also the way how nucleation is checked can be the culprit. Sometimes, there are enclosed liquid pockets, and you also need all those phases inside (sometimes even the primary fcc phase may be missing somewhere).
Another reason may be that a phase is existing, but interface mobility is too low.
In many cases, the last rests (e.g. <0.01%) of liquid are very difficult to remove, especially if diffusion in the solid phases is slow. Then it may be a pragmatic solution just to "switch" this rest to fcc once a sufficiently low temperature is reached. This can be done using the "add_to_grain" option with nucleation.
Bernd