hello everyone, I am trying to study the LAMMPS and hope to use lammps simulate a traditional triaxial tests. This means that I firstly need reach a isotropic compression, followed by a drained shearing simulations.
For the shearing part, it seems that fix deform is the best choice.
but for the isotropic compression part, there are many optionals.
I can use fix press/berendsen command to reach the target stress, and it works well, I assume that I can give a relatively large pdamp value to make this system stable.
But when I study the documentations of LAMMPS, it seems that “fix/relax command” can also be used to get the desired stress of the system. Besides, I can see that Fix nvt together with fix deform can also be used to reach the target stress.
it seems that all these methods are trying to change the dimension of the system so that the target stress is achieved. I am just wondering, is there any difference between these three methods?
or i can select any of them to get my target?
hope you have a nice day
following the previous concerns, how can we make sure that the system reached equilibration at the end of simulation
Both questions (which method to use and when equilibrium is reached) are not really questions about LAMMPS but questions about how to do your research and how to do MD simulations.
Thus whether to use, for example, fix press/berendsen + fix nvt or fix press/berendsen + fix nve + fix langevin or fix npt is a choice that you would have to make since all three combinations implement a barostat and a thermostat. There are more choices for the latter. If you want to make an educated choice, you have to educate yourself by studying the published literature discussing barostats and thermostats under special consideration of the kind of system you want to study. LAMMPS provides the tools, but you are doing the research, so you have to choose.
Fix box/relax is not suitable since already the first sentence of its documentation says that it is for use with energy minimization and not MD. This indicates that you are not using sufficient care when reading documentation or equivalent, so that is something you need to work on. How to improve your research skills is obviously something you need to discuss with your adviser/supervisor and not necessarily with people in an online forum. We are not your advisers! When you have to explain your choices later when trying to publish your results, you cannot write “some dude on the internet told me to do this”.
Using fix deform + fix nvt is obviously a very different method than using a barostat + a thermostat. You are setting how to deform your box in the first and let the system relax to match a specific pressure in the second. Thus, again, this is a choice you have to make when deciding how to study what you want to study. Each have their application. The best way to learn about this is, again, the published literature and the ultimate choice is yours and what to do a topic for a discussion with your adviser/supervisor.
Whether a system has reached equilibrium or not is a question that has plagued people doing MD simulations since the very beginning. There are no simple yes/no indicators. But, as before, this is not really a LAMMPS issue, but an issue of learning the MD simulation craft which should be a concern for your adviser/supervisor and less so for a discussion in a forum about an MD simulation software. At least not until somebody finds a reliable method to detect equilibrium that can be programmed as an option into a command like fix halt, of course .
thanks for your response and i will try to choose the suitable one based on research target.