Hello everyone,
I am relatively new to MICRESS(6.402) and had been trying to learn from the forums and the documentation.
I tried to simulate the solidification process of CuNiSn alloy, which involved the generation of two solid phases. I tried to obtain GES file from Thermal-calc2022 to provide part of the diffusion data and set interface seeds to fill the dendrite gap which was not filled by the first solid phase dendrites. However, the following error was reported during the operation of MICRESS. Moreover, the growth of the second solid phase is not ideal (there are many unstable interfaces), could you please tell me how to make the second solid phase grow more stably and make the software not report errors?
Best,
A ZU
Solidification of Cu alloys
Solidification of Cu alloys
- Attachments
-
- V~RP{PTP~~P$GKRH_F({%[V.png (109.13 KiB) Viewed 8864 times
Re: Solidification of Cu alloys
Dear A ZU,
Welcome to the MICRESS forum.
Unfortunately, there are various reasons which can lead to a behavior like what you describe, especially in an old MICRESS version like 6.4. But, generally, I would proceed in the following way:
1.) Check whether your secondary phase has been attributed to the correct thermodynamic phase. Furthermore, is it possible that this phase forms different composition sets (even if those are not included in the .GES5 file)? Then it may be important to set suitable start compositions when creating the .ges5-file.
2.) Have a look at the sub-lattices and solubilities of your secondary phase, which are written to the .log-file at time of initialization. You can see here the "Start Composition for iteration of quasi-equilibrium", in newer MICRESS versions you have also the solubility ranges. This tells you whether you have elements with restricted solubility (stoichiometric components), and whether the default major constituents have been correctly chosen when creating the .GES5-file (in case of composition sets).
3.) Have a look at the initial linearization parameters for the interaction of your secondary phase with liquid. These you also find in the .log-file, either at the beginning, or when first checking for nucleation of this phase was done. This initial equilibrium should be reasonable (correct compositions, slopes should not be tiny or extremely large). If this is not the case, or if you even get errors here, the problem must be fixed here (at time of initialisation).
4.) Check for error messages which can give information about specific failure (see here for more information).
5.) Check numerical parameters (too large interface mobility (if not "mob_corr"), strongly disbalanced interface energies, missing diffusion coefficients in liquid phase, ...), especially with respect to the phase interactions of your secondary phase.
6.) Check whether the composition of the liquid phase is negative for some of the elements (you should create phase composition outputs (.c*pha0) for that, preferably in at%) . Negative compositions can cause multiple problems. To cure them, use either more frequent updating of thermodynamic data or "penalty" terms on the driving force.
Which is the phase which creates the problems in your case, and which database do you use?
Best wishes
Bernd
Welcome to the MICRESS forum.
Unfortunately, there are various reasons which can lead to a behavior like what you describe, especially in an old MICRESS version like 6.4. But, generally, I would proceed in the following way:
1.) Check whether your secondary phase has been attributed to the correct thermodynamic phase. Furthermore, is it possible that this phase forms different composition sets (even if those are not included in the .GES5 file)? Then it may be important to set suitable start compositions when creating the .ges5-file.
2.) Have a look at the sub-lattices and solubilities of your secondary phase, which are written to the .log-file at time of initialization. You can see here the "Start Composition for iteration of quasi-equilibrium", in newer MICRESS versions you have also the solubility ranges. This tells you whether you have elements with restricted solubility (stoichiometric components), and whether the default major constituents have been correctly chosen when creating the .GES5-file (in case of composition sets).
3.) Have a look at the initial linearization parameters for the interaction of your secondary phase with liquid. These you also find in the .log-file, either at the beginning, or when first checking for nucleation of this phase was done. This initial equilibrium should be reasonable (correct compositions, slopes should not be tiny or extremely large). If this is not the case, or if you even get errors here, the problem must be fixed here (at time of initialisation).
4.) Check for error messages which can give information about specific failure (see here for more information).
5.) Check numerical parameters (too large interface mobility (if not "mob_corr"), strongly disbalanced interface energies, missing diffusion coefficients in liquid phase, ...), especially with respect to the phase interactions of your secondary phase.
6.) Check whether the composition of the liquid phase is negative for some of the elements (you should create phase composition outputs (.c*pha0) for that, preferably in at%) . Negative compositions can cause multiple problems. To cure them, use either more frequent updating of thermodynamic data or "penalty" terms on the driving force.
Which is the phase which creates the problems in your case, and which database do you use?
Best wishes
Bernd
Re: Solidification of Cu alloys
Dear Bernd:
I am also a beginner in MICRESS（6.402）, and currently my learning of MICRESS is limited to manuals and forums.I encountered the problem of negative concentration values while learning the MICRESS software. I simulated an Al-Si-Fe-Mg-Cr system, where the concentration of Cr appeared negative in the Cr deficient region. I have seen your suggestion and have conducted more frequent thermodynamic database updates (from 0.1s to 0.01s), but there are still negative values present.But when punishing the driving force, I am not sure where to make changes. Based on the prompts in the study manual, I believe that the Max value corresponds to the punishment you mentioned. I tried changing the Max value from 100 to 50,but there was still no significant effect (I tried to change Max in the interaction parameters of 0/1, 0/2, 0/3, and 0/4 respectively). But when Max reached 20, the matrix phase couldn't grow, and I tried to change the interface energy, but still couldn't grow. Is my modification of the parameters the same as your preset? This issue has been bothering me for a long time, and I urgently hope to receive your guidance.
Best，
Worker
I am also a beginner in MICRESS（6.402）, and currently my learning of MICRESS is limited to manuals and forums.I encountered the problem of negative concentration values while learning the MICRESS software. I simulated an Al-Si-Fe-Mg-Cr system, where the concentration of Cr appeared negative in the Cr deficient region. I have seen your suggestion and have conducted more frequent thermodynamic database updates (from 0.1s to 0.01s), but there are still negative values present.But when punishing the driving force, I am not sure where to make changes. Based on the prompts in the study manual, I believe that the Max value corresponds to the punishment you mentioned. I tried changing the Max value from 100 to 50,but there was still no significant effect (I tried to change Max in the interaction parameters of 0/1, 0/2, 0/3, and 0/4 respectively). But when Max reached 20, the matrix phase couldn't grow, and I tried to change the interface energy, but still couldn't grow. Is my modification of the parameters the same as your preset? This issue has been bothering me for a long time, and I urgently hope to receive your guidance.
Best，
Worker
Re: Solidification of Cu alloys
Dear worker,
Welcome to the MICRESS Forum!
While in newer MICRESS versions, the "penalty" function (which uses the modification of the driving force to prevent negative phase compositions) can be used independently from the "limits" (which do not allow TQ-equilibria outside a user-defined composition range for a given phase), in Version 6.4 they were still linked: As soon as the user defines a phase composition limit by using "limits" which restricts the allowed range further (as compared to the automatically detected solubility limits), penalty terms are automatically activated.
For your case (and MICRESS 6.402) the first step is to find out in which phase the composition is negative. For that aim you should have a look at the phase composition outputs (.c*pha*) if it is not obvious anyway. As you fight with negative compositions, you need to activate a penalty (by using "limits") for Cr in that phase by increasing the Min value to a small value above 0 (say 1.E-8 at%). Then, as soon as the composition of Cr in the specified phase falls below 1.E-8 at%, a penalty term on the driving force will trigger the phase transformation such that the Cr composition increases.
In newer MICRESS versions (from 7.0), the "penalty" function can be chosen independently from "limits" and can also be restricted to a specified phase interaction, which makes it more useful and reduces the probability of side effects.
Bernd
Welcome to the MICRESS Forum!
While in newer MICRESS versions, the "penalty" function (which uses the modification of the driving force to prevent negative phase compositions) can be used independently from the "limits" (which do not allow TQ-equilibria outside a user-defined composition range for a given phase), in Version 6.4 they were still linked: As soon as the user defines a phase composition limit by using "limits" which restricts the allowed range further (as compared to the automatically detected solubility limits), penalty terms are automatically activated.
For your case (and MICRESS 6.402) the first step is to find out in which phase the composition is negative. For that aim you should have a look at the phase composition outputs (.c*pha*) if it is not obvious anyway. As you fight with negative compositions, you need to activate a penalty (by using "limits") for Cr in that phase by increasing the Min value to a small value above 0 (say 1.E-8 at%). Then, as soon as the composition of Cr in the specified phase falls below 1.E-8 at%, a penalty term on the driving force will trigger the phase transformation such that the Cr composition increases.
In newer MICRESS versions (from 7.0), the "penalty" function can be chosen independently from "limits" and can also be restricted to a specified phase interaction, which makes it more useful and reduces the probability of side effects.
Bernd