I was trying to edit the crack example of the LAMMPS, which has been written for a 2D crack simulation. First of all, I am trying to change that to simulate a 3D crack. I am not sure that how should I change the s s p boundary conditions. Is that still logical for a simple 3D crack simulation?
Secondly, before the main simulation part, I am trying to add a equilibration part, in which by using NVT ensemble, I am trying the equilibrate the system. So at the end of this part the atoms have a velocity. However, in the main simulation I am adding a velocity to them. Won’t this sudden change make the system out of equilibrium?
Also, I know the Monte Carlo works based on the energy change. So in the example, the NVE fix is being used, in which the energy of the system is remained constant. I use fix atom/swap. Isn’t that paradoxical?
I am only commenting on that last point: LAMMPS will perform hybrid MC/MD simulation where atoms dynamics is controlled by MD (fix nve), and at intervals MC swaps are performed. The resulting simulation will not be in the NVE ensemble.
You should decide on that for yourself. It is good practice to think such things through.
What is “them”? You can only get meaningful answers on precise questions.
Please analyze very carefully what the crack example does.
No, it only means that you have not read the documentation carefully. Fix nve does not automatically mean that you have an NVE ensemble. All it does is plain time integration without any modification that changes temperature or pressure or number of particles. You only get an NVE ensemble, if you have a periodic bulk system and nothing else that changes things kinetic energy, volume, number of particles, or identity of particles. In the same way fix nvt only gives you an NVT ensemble if there are no other changes and so on.
OK. that’s great. I faced lost atoms error after using fix atom/swap and a change in its parameters could really change that when the lost error occurred. Do you have any idea idea that what could result to this error? Since there is no error without using fix atom/swap
This is a completely different topic than what was indicated in your original subject line. Please have the courtesy and post this as a new topic. This is an online forum and not personal emails, so you should keep things organized for the sake of the people that would be able to help you. But they can only do so if a) the topic is appropriate for the discussion and b) there is sufficient information to assess your problem.
This is not a case where you can look up what to do elsewhere, but where you have to carefully think about it yourself. Once you properly understand what kind of system you are simulating and for what purpose the different boundary conditions are chosen, the answer will be self-evident. An important step in research is that you do things for the right reason and not only because somebody says so.
That’s completely right. I just wanted some resources to find out what exactly each boundary condition means and being used for. Since knowing these correctly is what I’ll need to find out which one is appropriate for my system.
Have you ever thought about looking this up in the manual, for starters?
That said, I am tired of discussing at this level. You are making this far too much effort for very little gain. Please don’t expect any more responses from me to any of your posts.