LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] Simulation of polymer migration in cavities
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] Simulation of polymer migration in cavities

From: Axel Kohlmeyer <akohlmey@...24...>
Date: Tue, 20 Jun 2017 10:23:45 -0400

On Tue, Jun 20, 2017 at 8:45 AM, Nagarajan Karthik <a0108483@...33....6748...> wrote:

Dear all,


I wish to perform a Dissipative Particle Dynamics simulation of polymer chains migrating in an array of cavities in the presence of a steady external field. To do this, I am using a cubic simulation box of side length 20 with periodic boundary conditions in all 3 directions. There are 3 types of particles in the simulation. The solid mass comprises of 8818 particles arranged to form a spherical cavity. Also, there are 30 polymer chains with 10 beads each and 12481 solvent beads.  Please find attached the LAMMPS input script (‘ and data file (‘input_lam.dat’). When this simulation is run for more than 20 million time-steps with dt = 0.04, the component of the end-to-end vector of some chains along the direction of the field suddenly becomes very large (of the order of 10^4) at certain time-steps before returning to normal. I have attached a graph showing the variation in end-to-end vector of two chains (chains 14 and 24) with time. I am using chains with harmonic bonds. The force constant is 40 and the equilibrium bond length is 0.7. The problem does not arise in the absence of the external force.


I would appreciate if someone could clarify the following issues:


1.       What could be the cause of this problem? I have checked the values of various thermodynamic parameters (temperature, pressure etc) and nothing seems amiss. Also, why didn’t the simulation stop with a ‘Bond/Angle atom missing on proc ‘ error message when the chains became abnormally large?

your sys​tem will exhibit a drift in x-direction due to the added force representing the external field. this will lead to your entire system passing through perdiodic boundaries repeatedly. if you run for a very long time this will overflow the image flag counters (which are limited to 10 bit signed integers).
you don't get a bond missing error, because LAMMPS will look for the closest periodic neighbor, without consulting the image flags.


2.       What can be done to rectify this problem?

​you can add a restraining force​ to the part of the system that should remain stationary, e.g. using fix spring with the tether option, if you want to get rid of the drift.

if you prefer to keep the drift (remember, that the kinetic energy of the drift is included in your temperature), you have to re-compile LAMMPS with -DLAMMPS_BIGBIG so that image counters are stored in 21-bits signed integers instead.












Check out the vibrant tech community on one of the world's most
engaging tech sites,!
lammps-users mailing list

Dr. Axel Kohlmeyer  akohlmey@...12...24...
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.