About thermal coductivity calculation by Muller plathe method.

Dear Steve, and all users
I'm sorry for being late :frowning: And always, Thank you for your kind answers.

For your answer, Yes. It works when I didn't touch the Nevery value,
but doesn't when I change the Nrepeat value in the 'fix ave/time' even though I didn't do against a rule
(Nfreq must be a multiple of Nevery and Nevery must be non-zero even if Nrepeat is 1. Also, the timesteps contributing to the average value cannot overlap, i.e. Nrepeat*Nevery can not exceed Nfreq.)

Why is this happened....?
I think the Nrepeat value is important factor for calculation. What should I do?

Always, Thank you for your kindness.

From. Soon-sung

--------- ์›๋ณธ ๋ฉ”์ผ ---------๋ณด๋‚ธ์‚ฌ๋žŒ : Steve Plimpton <[email protected]>
๋ฐ›๋Š”์‚ฌ๋žŒ : "์†Œ์ˆœ์„ฑ" <[email protected]>
์ฐธ์กฐ : LAMMPS <[email protected]>
๋ฐ›์€๋‚ ์งœ : 2017๋…„ Mar 3์ผ(Fri) 02:32:47
์ œ๋ชฉ : Re: [lammps-users] About thermal coductivity calculation by Muller plathe method.
As the error message indicates, some fixes only
produce values that can be accessed at specific timesteps.

If you did this: "changed the values a little bit", you probably
changed some time frequencies and thus broke the script.

I assume the script worked w/out your changes?

Steve

Look at these 3 lines:

fix 2 all ave/chunk 2 500000 1000000 layers v_temp file profile.mp
variable Tdiff equal f_2[11][3]-f_2[1][3]
fix ave all ave/time 2 500000 1000000 v_Tdiff ave running

The last fix is evaluating the variable every 2 steps.
The variable references a fix that only produces its
output on steps that are a multiple of 1 million.

The error is telling you exactly that: the variable is using
a fix that isnโ€™t computed at a compatible time (i.e. the
timestep that the variable is being invoked on).

Steve

Dear all users

Always, Thank you for your kindness and answers and today I have a question.

For calculation of thermal conductivity of single layer graphene, we approximately choose the thickness to 0.34nm.

Once I calculated the single layer graphene with green-kubo relation, tersoff potential, and thickness of bigger than 0.34nm,
the values of thermal conductivity calculated are smaller (~150, 200W/Km) than the known ones ( 1000~3000W/Km).

Here are questions~!

1. I wonder why we have to use that value instead of bigger one.

2. Is it related to the Van der Waals Radius?

3. Is there any paper that can help me to understand it or I can refer to?

Plz give me a hand.
Again, thank you for reading my long question.

Dear all users
Always, Thank you for your kindness and answers and today I have a question.

For calculation of thermal conductivity of single layer graphene, we approximately choose the thickness to 0.34nm.
Once I calculated the single layer graphene with green-kubo relation, tersoff potential, and thickness of bigger than 0.34nm,the values of thermal conductivity calculated are smaller (~150, 200W/Km) than the known ones ( 1000~3000W/Km).

Here are questions~!
1. I wonder why we have to use that value instead of bigger one.

2. Is it related to the Van der Waals Radius?

3. Is there any paper that can help me to understand it or I can refer to?

Plz give me a hand.
Again, thank you for reading my long question.

The Green-Kubo relation is based on the autocorrelation function of heat flux tensors for a homogenous system. Since your system is inhomogeneous (with atoms and vacuum), Green-Kubo isnโ€™t the best approach for you. Try rNEMD (Muller-Plathe) method instead.

Ray

Thank you Ray~! It helps me a lot!
Again, Thank you for your kind answer~! :slight_smile:

Soon-sung.

--------- ์›๋ณธ ๋ฉ”์ผ ---------๋ณด๋‚ธ์‚ฌ๋žŒ : Ray Shan <[email protected]>
๋ฐ›๋Š”์‚ฌ๋žŒ : "์†Œ์ˆœ์„ฑ" <[email protected]>
์ฐธ์กฐ : lammps-mailinglist <[email protected]>
๋ฐ›์€๋‚ ์งœ : 2018๋…„ Oct 2์ผ(Tue) 22:08:38
์ œ๋ชฉ : Re: [lammps-users] Determination of thickness of Single-layer graphene in calculation of thermal conductivity
The Green-Kubo relation is based on the autocorrelation function of heat flux tensors for a homogenous system. Since your system is inhomogeneous (with atoms and vacuum), Green-Kubo isnโ€™t the best approach for you. Try rNEMD (Muller-Plathe)method instead.

Ray