If anyone has interest, I have written a new IO submodel, trackIO. This submodel outputs files and directory structures so that immediately following a run the particleTracks post-procsessing program in OpenFoam can be used to create VTK particle tracks. trackIO creates the time folders directly in the CFD folder. You can reconstructPar into these folders as usual. Note that I added one subprogram in IOmodel. An example snapshot of the VTK tracks for the Ergun test is included.
I have also written a short code, writeUfluid, to write the fluid velocity onto each particle in the langrangian directory. As a side benefit writeUfluid also re-writes the "positions" file with the correct cell IDs.
Mark
| Attachment | Size |
|---|---|
 newIOstuff.tar_.gz | 206.02 KB |
cgoniva | Sat, 07/28/2012 - 11:26
Hi Mark, I managed to include
Hi Mark,
I managed to include the code to the framework and will release it soon - thx for your contribution!!!. (some changes in writeUfluid were necessary...)
Do you know how to handle the particleTracks utility? I can only make it write on set of tracks using a given number of data points. The samplingFrequency setting does not work for me ???
Cheers,
Chris
schmeeck | Tue, 07/31/2012 - 16:12
Sorry, setting the sample
Sorry, setting the sample frequency to anything other than one is not working for me either. particleTracks only generates one set of tracks, as you mention. It seems as though you are aware that the -time option allows you to choose the beginning and end time frame of the tracks.
Thanks
Mark
cgoniva | Wed, 08/01/2012 - 14:42
Hi Mark, I see - so you made
Hi Mark,
I see - so you made the same experience as i did ;-)
Did you already test the writeUfluid and trackIO functionality from the latest release? I hope that I understood your code corrctly and included it in "your spirit"? Please let me know if any modifications sould be made.
BTW: don't you think it would now be "straight forward" to write the data to the processor directories and have it reconstructed with the other data? The writing should then only be done if the particle is on the processor domain, which is true, when cellI >-1.
Cheers,
Chris
schmeeck | Fri, 08/10/2012 - 22:47
Thanks Chris. I have been
Thanks Chris. I have been using your modified trackIO and writeUfluid. I don't know why I was writing uFluid before it had been read. Lack of sleep I guess. I have written some code that calculates particle-fluid correlations using uFluid and have been getting very interesting results. Thanks again.
Your solution to write to the processor directories certainly makes sense.