Dumping Velocity

[abbaspour...@gmail.com]

Hello everyone,

I am running a simulation in Hoomd-blue, using  Brownian dynamic simulation for polymer chains  without thermal noise.,"integrator=hoomd.md.integrate.brownian(group=polymer, kT =1.0, seed=100, noiseless_r=True, noiseless_t =False)"

I want to dump particle position along with their velocities. for that, I'm using "hoomd.dump.gsd("trajectory.gsd", period=500, group=all, dynamic=['momentum'], overwrite=True) .

There is an active force in the system which makes the polymer chain have a bias direction. If I visualize the simulation on Ovito I can clearly see it has more directed motion but when I look at the velocity profile there is a lot of noise there which looks it has more random motion. what could be the reason?

Bests,

Leila

[Ehsan Irani]

Dear leila,

Using brownian integrator makes velocities decouple from positions, so they are drawn from a distribution at each timestep randomly. You need to calculate delta_r/dt yourself to find the "real" velocities.

Best,

Ehsan

--
You received this message because you are subscribed to the Google Groups "hoomd-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hoomd-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hoomd-users/53df0650-7b28-4fa8-8221-f1f3f8a1d367n%40googlegroups.com.

[Leila Abbaspour]

Thanks Ehsan for your prompt response. 

To view this discussion on the web visit https://groups.google.com/d/msgid/hoomd-users/CAFw_vxXhoU-ZQ0ncwmZFGfhR9QTkf2Xgfg6uESyDLQpnhjUNqw%40mail.gmail.com.

[Joshua Anderson]

All,

This has been a really common confusion regarding the Brownian Dynamics integrator. Various parts of the codebase require that the particle velocity field be consistent with the equations of motion in the system, so I am not going to change that.

Please correct me if I am wrong, but I think what you are interested in is a measure of the difference in position for a particle at two different points in time \vec{r}_i(t) - \vec{r}_i(t-1), taking into account periodic boundary conditions.

Would someone be willing to add this functionality?

In HOOMD v3, you could add an object that stores the positions at a time lag and provides a loggable quantity for the difference in positions. You could log this quantity to a GSD file. The advantage of this method over using consecutive frames in the GSD file is that you would be able to use a shorter lag for the delta (such as 1) than the number of timesteps between frames in the trajectory.
------
Joshua A. Anderson, Ph.D.
Research Area Specialist, Chemical Engineering, University of Michigan

> To view this discussion on the web visit https://groups.google.com/d/msgid/hoomd-users/CABOSGCtptf_jL8X3amzO6d1-2%3DAtWC9inndSj7aK6L5bv%3D_-8g%40mail.gmail.com.

[Joshua Anderson]

All,

Dan Evans added a new integrator that implements overdamped viscous motion - consistent with hoomd v2 `brownian` with noiseless_t=True and noiseless_r=True. Without the thermalizing Brownian motion applied, OverdampedViscous sets the velocities to the force divided by the drag coefficient (and similarly for angular momentum).
https://github.com/glotzerlab/hoomd-blue/pull/1121

This integration method will be available in the next HOOMD v3.0 beta release. This provides the velocity logging that many of you have been asking for in a way that is consistent with the documented equations of motion.

------
Joshua A. Anderson, Ph.D.
Research Area Specialist, Chemical Engineering, University of Michigan