Random-walk diffusion model

Submitted by rqwang on
ckloss

>>My guess is at LIGGGHTS source code
Correct. I see two possibilities:

A) change the integration in order to do this. This would be an extension or a derived class of FixNveSphere, see fix_nve_sphere.h/cpp

B) you could also apply a random force which should be equivalent. This would be a modified fix addforce, see fix_addforce.h/cpp

Christoph

Thu, 3 Nov 2011 - 17:37 Permalink
cgoniva

In reply to by anonymous_stub (not verified)

Hi!

Alternatively you could think about some random forces on the particles - if time resolution of the coupling is high enough you could do that via a simple force model (coupled to turbulence parameters...).

Cheers,
Chris

Fri, 4 Nov 2011 - 09:15 Permalink

Hi,

I am trying to add an artificial random-walk diffusion model.

It has this form:

xp(t + ∆t ) = xp(t ) + up ∆t + ξ sqrt(2 Ds ∆t)

where Ds is a constant, ξ is a random number from 0 to 1.

Could anyone tell me where I can implement this?
My guess is at LIGGGHTS source code, but I have no idea which file to modify.

Please help me!
Thanks.

rq