Re: [ESPResSo-users] Any output for tutorial5-1.tcl (streamline movie)??

[Georg Rempfer]

Hello Pet,

I don't know what you mean by "I am sure of its parameters", but I
know that in an LB fluid of density 1 and temperature 1 fluctuates
like crazy. Also this is exactly the way it should be, as on the
nanoscale for a system at room temperature, the fluctuations are
orders of magnitude bigger than any drift velocity.

If you want nice videos like the ones I sent you, you will have to
switch the thermalization off completely. You can achieve this with
"thermostat off" instead of "thermostat lb 1.". However, if you run a
coupled MD LB simulation without fluctuations, you are essentially
simulating a system at 0 degrees kelvin. So depending on what you want
to do, this would be terribly wrong.

If you want to have a physically reasonable simulation _and_ smooth
streamlines, you will have to do what Ulf Schiller says, which is
average a lot of frames. The fluctuations will average out to 0, all
you will see is the drift. If you have a non zero drift velocity,
however, your particles move around and if you are interested in the
velocity field around some macro molecule, you would have to do the
averaging in a co-moving reference frame, which is only possible if
you have one molecule.

Regards,
Georg

2013/9/7 pdam bright <address@hidden>:
> Dear all,
> As I mentioned in my previous e-mail, when I plot the .vtk output of my
> simple code, the streamlines aren't reasonable and they are independent
> frames just come after each other. One of the answers to my question was
> that maybe the parameters in my code produce large fluctuations.
>
> So I decided to run the sample code "tutorial5-1.tcl", which I am sure of
> its parameters, to see if my procedure in paraview will produce a good
> streamline movie or not? But I still couldn't use it properly and from one
> frame to the next one there is a big jump in streamlines!
>
> Can anyone provide me with a good streamline animation from this sample
> code, I've attached the code in the following, I've only added this
>     set out "field_$i.vtk"
>     lbfluid print vtk velocity $out
> in 3# place to make .vtk files.
>
> Thanks in advance,
> Pet
>
>
> ######################################################
> #
> # Diffusion of a single particle with hydrodynamic interactions
> (Lattice-Boltzmann)
> #
> #########################################################################################
> ## set up the box
> setmd box_l 16. 16. 16.
> ## the skin is not important for a single particles
> setmd skin 0.2
> ## the MD step in this simulation could be bigger, but
> ## we keep it consistent with typical LJ simulations
> setmd time_step 0.01
> ## LB can not handle Verlet lists, so we disable them
> cellsystem domain_decomposition -no_verlet_list
>
> ## open files for output
> set posfile [ open "pos.dat" "w" ]
> set vfile [ open "vel.dat" "w" ]
> set enfile [ open "energy.dat"  "w"]
>
> ## this creates the LB fluid with the parameters
> ## of our choice.
> lbfluid grid 1. dens 1. visc 1. tau 0.1 friction 50.
> ## this activates the thermalization of the
> ## LB fluid and adds the random forces on the
> ## particle
> thermostat lb 1.
>
> ## create a particle
> part 0 pos 0 0 0
>
> set out "field_0.vtk"
> lbfluid print vtk velocity $out
>
> ## perform a couple of steps to come to equilbrium
> puts "Warming up the system."
> integrate 1000
> puts "Done."
>
>
> set out "field_1.vtk"
> lbfluid print vtk velocity $out
>
> ## Now the production run!
> for { set i 2 } { $i < 1000 } { incr i } {
>   if { $i % 100 == 99 } {
>     ## tell the user we are still working
>     puts "cycle $i completed"
>   }
>
>   integrate 10
>
>     set out "field_$i.vtk"
>     lbfluid print vtk velocity $out
>
>   puts $posfile "[ setmd time ] [ part 0 print pos ]"
>   puts $vfile "[ setmd time ]  [ part 0 print v ]"
>   puts $enfile "[ setmd time ]  [ analyze energy kinetic ]"
> }
>
> ## and close output files
> close $posfile
> close $vfile
> close $enfile