Re: [ESPResSo-users] Regarding Standard DPD vs Transverse DPD

[Christoph Junghans]

Hi Salvador,

Am 10. März 2012 16:06 schrieb Salvador H-V <address@hidden>:
> I am perfuming some simulations using DPD in ESPReSo to reproduce the
> coefficient diffusions reported in: Soft Matter 2008,4,156.
We have done this study with a modified version of 2.0.3o(?) and a lot
of things have changed since then.
>
> In particular i am trying to reproduce the diffusion coefficient of 4000 LJ
> repulsive (WCA) particles in 3D (figure1).
>
> I create a script with the specified conditions and parameters by the
> authors. Then I compiled ESPResSo twice, one with DPD feature activated and
> another with only Trans_DPD.
>
> Here it is the code info for each compilation:
>
> a) Just DPD feature
> ESPResSo: 3.0.2
> { Compilation status { FFTW } { MODES } { PARTIAL_PERIODIC } {
> ELECTROSTATICS } { MASS } { EXTERNAL_FORCES } { CONSTRAINTS } {
> BOND_CONSTRAINT } { EXCLUSIONS } { TABULATED } { LENNARD_JONES } {
> SOFT_SPHERE } { LJCOS } { LJCOS2 } { BOND_ANGLE_HARMONIC } { NPT } { DPD } }
> { Debug status { MPI_CORE FORCE_CORE } }
>
> b) Trans_DPD
> ESPResSo: 3.0.2
> { Compilation status { FFTW } { MODES } { PARTIAL_PERIODIC } {
> ELECTROSTATICS } { MASS } { EXTERNAL_FORCES } { CONSTRAINTS } {
> BOND_CONSTRAINT } { EXCLUSIONS } { TABULATED } { LENNARD_JONES } {
> SOFT_SPHERE } { LJCOS } { LJCOS2 } { BOND_ANGLE_HARMONIC } { NPT } {
> TRANS_DPD } { DPD } }
> { Debug status { MPI_CORE FORCE_CORE } }
>
> To run the simulations I am using for both kind of simulations (thermostats)
> the following commands:
>
> thermosat off
> galileiTransformParticles
> set temp_ave 1.2
> set gamma_dpd 1.0
> set dpd_cut [ expr { ( pow ( 2.0 , 1.0/6.0 ) )*$diam } ]
> thermostat dpd $temp_ave $gamma_dpd $dpd_cut WF 1
looks good so far.
>
> to obtain the different diffusion coefficients I have changed gamma_dpd
> accordingly to {0.5,1.0,2.0,3.0,5.0,7.0 & 10.0}
>
> For the kinetic temperature and pressure I am obtaining similar values as
> those reported in the article for both thermostats.
>
> Using the standard DPD thermostat I am able to obtain a similar behavior of
> Dt as the authors shown in figure1. However, even when I am using the same
> script and running the simulation with the correspondingly ESPResSo
> compilation I can not reproduce the results for the coefficient diffusion
> using the Transverse-DPD thermostat. I mean, I can not observe the
> drastically reduction in the coefficient diffusion as gamma_dpd increases
> form 0.5 to 10.
>
> Anny comment or suggestion of possible causes of errors or mistakes in my
> procedure would be greatly appreciated.
Do use [analyze MSD] to measure the diffusion constant?
Did you set gamma for the (non-trans) dpd part to 0 for the transverse
study and the trans WF to 1? (check output of [thermostat])?
Do your particles have mass 1?

If all this does not help, I can try to find the data on one of my old
harddrives.
(Please contact me off-list for that.)

Cheers,

Christoph

>
> Thanks a lot,
>
> Salvador
>



-- 
Christoph Junghans
Web: http://www.compphys.de