CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ParaView (https://www.cfd-online.com/Forums/paraview/)
-   -   [General] Stream Tracer Strange Behaviour (https://www.cfd-online.com/Forums/paraview/138342-stream-tracer-strange-behaviour.html)

Ric.E July 2, 2014 12:25

Stream Tracer Strange Behaviour
 
Hi,

I've noticed some strange behaviour when plotting stream tracers in an electric field vector field. The line integration doesn't seem to terminate at the mesh boundary, as is described in the documentation, the lines go off at random directions and there is some bunching of lines (attach .4).

I've searched the forums and found this post: http://www.cfd-online.com/Forums/par...mulations.html which I followed, but still get erroneous output. (attach. 1)

I also found something about stream tracers not really being meant for 2D models in case of errant z-components throwing the tracer off. I followed the guidelines for this by applying a calculator result assigning Ex to x_hat and Ey to y_hat, to restrict the result to the plane. I then did the streamer but saw no change.

I have also tried increasing the resolution of the mesh and the line integral step calculation with no improvement. attach. 2 shows a min, max and initial step calculation of 10e-5.

My file is vtu file, generated from ElmerFEM, so I believe the mesh should be included and it's not just an array of points (as with a CSV), so the mesh border should be visible to paraview (is this correct?). In case though, I tried adding a Delauney2D filter. This resulted in the holes of my 2D plane being filled in, which is not what I'm after (attach.3). However, the results do look more correct. perhaps this is the only option?

I also plotted the field vectors incase there was an error in the result file calculations.

I've included an example result file and the paraview save state, along with images named above, in a zip file:

https://dl.dropboxusercontent.com/u/...achment.tar.gz

If anyone can help at all or has any suggestions I would be very grateful!

Many Thanks
Richard


All times are GMT -4. The time now is 11:41.