ORE: Query on AtC command

Hello,Thanks again for your helpful assistance .However,I still can‘t underatsand some answers .Please give me more details about the questions below .I am very grateful!
Q6:In the example 5.2,the paper use the AI-REBO potential between the gold atoms and the GS atoms .I have read the “pair_style airebo” command in the manual.The command’s syntax is “pair_style airebo cutoff LJ_flag TORSION_flag”.However, the paper didn’t tell us what is the number of the cutoff in the command .So, I want to know what is the number of the cutoff .
Q7: As you said , The faceset obndy only include the leftmost face and the rightmost face in the simulation box (-12 12 -3 3 -3 3) due to periodicity condition ,i.e they are x=12 face and x=-12 face .Is right my understanding?
Q8:In general,there are two kinds of boundary conditions .They are Dirichlet boundary condition(temperature) and Neumann boundrary condition(heat flux),respectively.(1)So,which kind of boundary condition is used in the in_bar1dflux? Becasue I think that the command "fix_modify AtC transfer thermal control flux faceset obndy " defines the Neumann boundrary condition while the command “fix_modify AtC transfer fix temperature lbc 40、fix_modify AtC transfer fix temperature rbc 20” defines the Dirichlet boundary condition.(2) If the in_bar1dflux only defines the Dirichlet boundary condition as if what the in_bar1d have done,Can you tell me how to apply the Neumann boundrary condition(heat flux) to simulation system as if what the example problem(Templeton, Jones, & Wagner, MSMSE 18 2010) have done ?(3) If the in_bar1dflux only defines the Neumann boundrary condition(heat flux), which command is used to achieve the boundary condition?
Q9:If my understanding to your answer is right, I can’t agree with you. Because the commands “variable xdof equal 3*count(ghost)
compute_modify thermo_temp extra ${xdof}” have removed the contributes of ghost atoms ,then,Why the “temp” produced by thermo command is different from the atomic_temperature_mean in bar1d_fluxFE.GLOBALS?
Q12: Can you tell me how to judge if the MD filed is equilibrium by those metrics in more details?
Thank you very much !
Best regards!

在 2011年9月13日 上午12:34,Templeton, Jeremy Alan <jatempl@…3…>写道:

More below:

Q6:In the example 5.2,the paper use the AI-REBO potential between the gold atoms and the GS atoms .I have read the “pair_style airebo” command in the manual.The command’s syntax is “pair_style airebo cutoff LJ_flag TORSION_flag”.However, the paper didn’t tell us what is the number of the cutoff in the command .So, I want to know what is the number of the cutoff .

No, AIREBO is only used between GS atoms. The GS-gold interaction is LJ.

Q7: As you said , The faceset obndy only include the leftmost face and the rightmost face in the simulation box (-12 12 -3 3 -3 3) due to periodicity condition ,i.e they are x=12 face and x=-12 face .Is right my understanding?

Yes.

Q8:In general,there are two kinds of boundary conditions .They are Dirichlet boundary condition(temperature) and Neumann boundrary condition(heat flux),respectively.(1)So,which kind of boundary condition is used in the in_bar1dflux? Becasue I think that the command "fix_modify AtC transfer thermal control flux faceset obndy " defines the Neumann boundrary condition while the command “fix_modify AtC transfer fix temperature lbc 40、fix_modify AtC transfer fix temperature rbc 20” defines the Dirichlet boundary condition.(2) If the in_bar1dflux only defines the Dirichlet boundary condition as if what the in_bar1d have done,Can you tell me how to apply the Neumann boundrary condition(heat flux) to simulation system as if what the example problem(Templeton, Jones, & Wagner, MSMSE 18 2010) have done ?(3) If the in_bar1dflux only defines the Neumann boundrary condition(heat flux), which command is used to achieve the boundary condition?

This one it is important to stay clear on. There are boundary conditions for the system, both Dirichlet and Neumann, and then there is the MD-FE coupling, which has a similar flavors of heat flux coupling or temperature coupling. Both types of BCs can be applied to FE-only elements based on user specification no matter what. If BCs are applied to the MD system, then the control must be turned on, but again, both types of BCs can be applied. The syntax “flux faceset obndy” tells atc that the MD-FE coupling will be based on conserving the heat flux, and that the coupling flux is defined relative to the obndy faceset. in.bar1d_flux applies Dirichlet boundary conditions to the FE at each end while performing MD-FE coupling using the heat flux.

Q9:If my understanding to your answer is right, I can’t agree with you. Because the commands “variable xdof equal 3*count(ghost)
compute_modify thermo_temp extra ${xdof}” have removed the contributes of ghost atoms ,then,Why the “temp” produced by thermo command is different from the atomic_temperature_mean in bar1d_fluxFE.GLOBALS?

Sorry, I should have been more clear. Ghost atoms in atc are different that ghost atoms in lammps. Ghost atoms for atc are real lammps atoms but used to help define the MD region’s boundary and reflect phonons.

Q12: Can you tell me how to judge if the MD filed is equilibrium by those metrics in more details?

There are many others on this users list who are more knowledgeable than I. Typically what I do is output and rescale and different frequencies so that I can see how far away from the target temperature I am. You can also computed windowed averages which should match the desired statistical properties.
Jeremy

Hello, I am very grateful for your help.But I still have some quesetions about the answers.
Q6: In the example 5.2,the paper use the AI-REBO potential between the GS atoms .I have read the “pair_style airebo” command in the manual.The command’s syntax is “pair_style airebo cutoff LJ_flag TORSION_flag”.However, the paper didn’t tell us what is the number of the cutoff in the command .So, I want to know what is the number of the cutoff in the paper .
Q8:As you said,
(1) The in.bar1d_flux applys the Dirichlet boundary condition to ‘lbc’ and ‘rbc’ while it applys the Neumann boundrary condition to MD boundary. Is right my understanding?
(2) Both the “thermal control hoover"and the "thermal control flux" only be used to MD system.Is right my understanding? Does the ‘hoover’ fix temperature ? Does the ‘flux’ apply the Neumann boundrary condition to MD ? The two commands are goal to couple MD with FE ?
(3) In a simulation , How to choice the “hoover” or " flux"? Because I don’t know which one should be used when I want to implement a simulation**.**
(4) The command “fix_modify AtC transfer fix” applys the Dirichlet boundary condition to FE ,however,which command can be used to apply the Neumann boundrary condition to FE ? For example,in the example 5.1(Templeton, Jones, & Wagner, MSMSE 18 2010 P14), the heat flux=3.32*(10^9)w/m^2 is applied to z boundary ,Can you tell me how to achieve it ? In additional, Are the heat flux be applied to FE(not MD )?
(5) If you don’t mind, Can you give me the input stream in example 5.1 ? I will be very grateful !
Q9: I can’t understand your meaning . In fact ,the temp only calculate the atoms in group “internal” due to the command “variable xdof equal 3*count(ghost) compute_modify thermo_temp extra ${xdof}” ,then How is the atom_mean_temperature calculated? Why are they different ?
Best regards!

在 2011年9月13日 下午11:03,Templeton, Jeremy Alan <jatempl@…3…>写道:

More below:

Q6: In the example 5.2,the paper use the AI-REBO potential between the GS atoms .I have read the “pair_style airebo” command in the manual.The command’s syntax is “pair_style airebo cutoff LJ_flag TORSION_flag”.However, the paper didn’t tell us what is the number of the cutoff in the command .So, I want to know what is the number of the cutoff in the paper .

We used a scale factor of 1.5 with the standard implementation (see the doc page for more).

Q8:As you said,
(1) The in.bar1d_flux applys the Dirichlet boundary condition to ‘lbc’ and ‘rbc’ while it applys the Neumann boundrary condition to MD boundary. Is right my understanding?

Yes, but the heat flux coupling is not a boundary condition in the traditional sense. Rather it is an energy exchange between the two systems, and the FE equations are modified as well as per Wager et al.

(2) Both the “thermal control hoover"and the "thermal control flux" only be used to MD [system.Is](http://system.Is) right my understanding? Does the ‘hoover’ fix temperature ? Does the ‘flux’ apply the Neumann boundrary condition to MD ? The two commands are goal to couple MD with FE ?

Yes, your understanding is correct. Just to be clear, the fixed temperature vs. Neumann (as you describe them) only apply to FE/MD coupling. Either method is capable of applying a true boundary condition to MD as in Templeton et al.

(3) In a simulation , How to choice the “hoover” or " flux"? Because I don’t know which one should be used when I want to implement a simulation**.**

We suggest flux. After adding Dirichlet BCs in Templeton et al., we provide the option to do coupling based on it in the code. However, the flux coupling has some nice properties in terms of global energy conservation and not forcing the MD at the boundary to match the FE, which it shouldn’t.

(4) The command “fix_modify AtC transfer fix” applys the Dirichlet boundary condition to FE ,however,which command can be used to apply the Neumann boundrary condition to FE ? For example,in the example 5.1(Templeton, Jones, & Wagner, MSMSE 18 2010 P14), the heat flux=3.32*(10^9)w/m^2 is applied to z boundary ,Can you tell me how to achieve it ? In additional, Are the heat flux be applied to FE(not MD )?

In that paper, the heat flux is applied to the MD using the FE as a framework to help define geometry. The the command to apply the heat flux is documented on the Lammps site: http://lammps.sandia.gov/doc/USER/atc/man_fix_flux.html

(5) If you don’t mind, Can you give me the input stream in example 5.1 ? I will be very grateful !

There are two parts to 5.1, you’ll have to specify.

Q9: I can’t understand your meaning . In fact ,the temp only calculate the atoms in group “internal” due to the command “variable xdof equal 3*count(ghost) compute_modify thermo_temp extra ${xdof}” ,then How is the atom_mean_temperature calculated? Why are they different ?

Can you please send me the input deck and output you’re generating so I can take a look.
Jeremy