[Paraview] particle tracing and particle tails help

Dan Lussier dtlussier at gmail.com
Wed Apr 6 15:36:44 EDT 2011


Hi.

I am having a lot of trouble getting particle tracing to work - and  
frustratingly can't even get any output to work so trouble shooting is  
very difficult.  I've broken down my situation below:

Setup:
-------

* PV 3.11 from git
* pv-Meshless plugin installed

Data:
-----

* time-aware XDMF files
* particles represented in PV on unstructured mesh using XDMF  
TopologyType="Polyvertex"

* Usually 100,000 to 10 million particles
	- Reduced size sample files are available at:

http://dl.dropbox.com/u/17666990/sample_particle_data.zip

	- Each individual file is a single frame, which are assembled in a  
XDMF temporal collection in dump_argon_state_rdfval2_TEMPORAL- 
SUMMARY.xmf.  This is the file that should be loaded to allow for  
access to all frames.


Output
--------

I would like to be able to trace the path of a subset of particles  
from within the larger dataset over time, and have its path traced out  
in 3D space.  This is in contrast to tracing the advection of trace  
particles within a vector field, which is possible from what I  
understand via the ParticlePathlines filter.

I have tried playing around with the ParticleTracer and  
ParticlePathlines filters in PV, but cannot get anything to work.

One thing I'm wondering about is whether my particles jump around too  
much (i.e. their displacement is too large between frames).  It is  
likely that this occurs because each frame represents the output from  
simulation code which may be separated by between 100 to 10000  
simulation timesteps.

I have tried crawling through the PV mailing list archives but wasn't  
able to find many comments that were current, and fit what I was  
looking for.  I am open to any suggestions on how I could get this  
working using my data, and would really, really appreciate if someone  
could point me towards some examples online that show the actual  
pipelines with sample data to pull off something like this.

Thanks,

Dan



More information about the ParaView mailing list