[fengjq]

Hi,

Would some of you please share experience with using paraFoam to display Lagrangian particle tracing? I'm using OpenFOAM 2.2.2 and tried to compute the "sprayFoam" tutorial case "aachenBomb". The solver finished computing and generated data files. But I have no clue on how to get paraFoam to display the particles and their pathlines. Any comments on the general procedures for Lagrangian cases can also be of great help for a new OpenFOAM user like me. Thanks very much.

James

[ansubru]

Hey Foamer!!

I am a newbie just like you and I am using DPMFoam.. so particle tracking is something I have been doing. Export your case to paraview.. In the geometry view tree to the left (where different parts of your geometry are displayed).. select... kinematic cloud (or whichever cloud you have used.. could be colliding cloud etc.).. create a glyph.. while having selected the corresponding cloud.. here you choose the method of tracking you want.. solid spheres or path lines. arrows..etc.... click on apply.. and Voila!! you are done.. for more details refer earlier threads in this forum..

http://www.cfd-online.com/Forums/ope...-paraview.html

Hope this helps..

Cheers!!!

[fengjq]

Originally Posted by ansubru

Hey Foamer!!

I am a newbie just like you and I am using DPMFoam.. so particle tracking is something I have been doing. Export your case to paraview.. In the geometry view tree to the left (where different parts of your geometry are displayed).. select... kinematic cloud (or whichever cloud you have used.. could be colliding cloud etc.).. create a glyph.. while having selected the corresponding cloud.. here you choose the method of tracking you want.. solid spheres or path lines. arrows..etc.... click on apply.. and Voila!! you are done.. for more details refer earlier threads in this forum..

http://www.cfd-online.com/Forums/ope...-paraview.html

Hope this helps..

Cheers!!!

Thanks, Ananda. You are actually ahead of me. I'm not sure if I know how to "export your case to paraview ..." When "sprayFoam" finished computing I can see a directory "postProcessing/lagrangian/sprayCloud/..." appears in my case directory (which seems to contain the lagrangian particle data files at each time step). I also tried to do "foamToVTK" which generates another directory "VTK/lagrangian/sprayCloud/ ..." But when I opened paraview, I don't know which file I should open (trying to open the "sprayCloud ... vtk" or "aachenBomb...vtk" did not get anything displayed).
Any more detailed instructions? Thanks.

James

[wyldckat]

Quick answer: http://openfoamwiki.net/index.php/Ma...gian_particles

[ansubru]

Hey there!!

You need to open each and every subfolder within VTK, then select the object which has a +sign before it...(you dont need to click on the + sign as it would just open up the tree..)... selecting this would select all the data for each time step... load each of these VTK files, and select the kinematic sub section.. as i quoted earlier apply a glyph here.. and ur done..

[fengjq]

Originally Posted by ansubru

Hey Foamer!!

I am a newbie just like you and I am using DPMFoam.. so particle tracking is something I have been doing. Export your case to paraview.. In the geometry view tree to the left (where different parts of your geometry are displayed).. select... kinematic cloud (or whichever cloud you have used.. could be colliding cloud etc.).. create a glyph.. while having selected the corresponding cloud.. here you choose the method of tracking you want.. solid spheres or path lines. arrows..etc.... click on apply.. and Voila!! you are done.. for more details refer earlier threads in this forum..

http://www.cfd-online.com/Forums/ope...-paraview.html

Hope this helps..

Cheers!!!

Originally Posted by wyldckat

Quick answer: http://openfoamwiki.net/index.php/Ma...gian_particles

Thanks for pointing this link, which I looked before and following the steps I could not see "(Lagrangian) clouds" in "Mesh Parts". So I'm stuck there. I tried File > Open > VTK > ... vtk and couldn't get anything meaningful (the "Object Inspector" is all blank). Am I missing something else?

By the way, is there a general instruction on the steps for paraFoam with Lagrangian data for OpenFOAM 2.2.2 in particular (most of those postings were for earlier versions)?

[ansubru]

Hi Again!

Very interesting.. when you execute foamToVTK in ur terminal.. does openfoam execute some lines of code which are visible in ur window?? If so, the files have been correctly converted..

There are no general guidelines (at least as far as I am aware).. i myself am using OF 2.3.0 and its rather intuitive to implement particle tracking.. you just need to ensure that there are some VTK files being created.. corresponding to each time step..

I know some of these views are rather obtuse, but the idea is that you learn by fiddling around ..

Added tip, try to use paraview as a stand alone.. dont use parafoam.. At least thats what I did.. try to build ur own paraview.. with instructions in paraview home page.. and then it is sure to work..
(do this at your own risk, building paraview is not very straightforward by any means)....
Cheers!!!

[fengjq]

Originally Posted by ansubru

Hi Again!

Very interesting.. when you execute foamToVTK in ur terminal.. does openfoam execute some lines of code which are visible in ur window?? If so, the files have been correctly converted..

There are no general guidelines (at least as far as I am aware).. i myself am using OF 2.3.0 and its rather intuitive to implement particle tracking.. you just need to ensure that there are some VTK files being created.. corresponding to each time step..

I know some of these views are rather obtuse, but the idea is that you learn by fiddling around ..

Added tip, try to use paraview as a stand alone.. dont use parafoam.. At least thats what I did.. try to build ur own paraview.. with instructions in paraview home page.. and then it is sure to work..
(do this at your own risk, building paraview is not very straightforward by any means)....
Cheers!!!

When I do "foamToVTK" I indeed see OpenFOAM running lines on my screen and it finished with a new directory "VTK" put in my case directory which contains many (one for each time step) *.vtk files. So I believe the correct files were generated. But when I tried to open any of those .vtk files in paraview, the "Object Inspector" is all blank (where there is no "Mesh Parts", etc.) and in the display window all I see is the mesh.

I actually compiled paraview on my computer (when building the paraFoam). Are you saying that I should go to the individual subdirectories to build paraview?

By the way, I'm interested in simulating aerosol particles in gas stream in channels. The "sprayFoam" was what I could find in OpenFOAM Lagrangian tutorials that seems relevant (although more stuff than I need) among others. It sounded like there was an "icoLagrangianFoam" in earlier OF versions. I wonder if you get different LagrangianFoams in OF 2.3.0.

Thanks.

[ansubru]

Hi again!!

The way you compile paraview is totally upto you. Just ensure that it works. I compiled it using thirdparty (provided in the source build of OF..

PHP Code:

http://www.openfoam.org/download/source.php 


Yes OF 2.3.0 has icoLagrangianfoam. I have used it personally, and there is good tutorial (the hopper). So i guess you should look into that. The earlier versions of OF also had this solver. In case you are using a version which doesnt include the Icolagrangianfoam, you can alsways download it from the Git (as wyldckat suggests).

Regards

ansubru

[fengjq]

Hi ansubru,

I just installed OpenFOAM 2.3.0 and tried to follow your description. I went to the /tutorials/lagrangian/icoUncoupledKinematicParcelFoam/hopper directory, and did the following.

1. ./Allclean
2. cd hopperInitial , blockMesh, icoUncoupledKinematicParcelFoam (to get the solver running)
3. foamToVTK (after solver finished)
4. paraview, File > Open > VTK/hopperInitialState_..vtk ("OK", "Apply" so I see the hopper domain image displayed in the paraview window)
5. What's visible are "Properties (hopperInitialState_..*)", "Display (UnstructuredGridRepres", "Representation, Surface", "Coloring", etc. But I do not see the "geometry view" (what is it?) and "Mesh Part" (as what you described).

What should I do at this point? Please be patient with me. I think I need some more experienced OF user to lead me through the "baby-steps". Thanks.

fengjq

Originally Posted by ansubru

Hi again!!

The way you compile paraview is totally upto you. Just ensure that it works. I compiled it using thirdparty (provided in the source build of OF..

PHP Code:

http://www.openfoam.org/download/source.php 


Yes OF 2.3.0 has icoLagrangianfoam. I have used it personally, and there is good tutorial (the hopper). So i guess you should look into that. The earlier versions of OF also had this solver. In case you are using a version which doesnt include the Icolagrangianfoam, you can alsways download it from the Git (as wyldckat suggests).

Regards

ansubru

Thanks for the advice. I'll give it a try and let you know how it works.

[fengjq]

Hi ansubru,

I just installed OpenFOAM 2.3.0 and tried to follow your description. I went to the /tutorials/lagrangian/icoUncoupledKinematicParcelFoam/hopper directory, and did the following.

1. ./Allclean
2. cd hopperInitial , blockMesh, icoUncoupledKinematicParcelFoam (to get the solver running)
3. foamToVTK (after solver finished)
4. paraview, File > Open > VTK/hopperInitialState_..vtk ("OK", "Apply" so I see the hopper domain image displayed in the paraview window)
5. What's visible are "Properties (hopperInitialState_..*)", "Display (UnstructuredGridRepres", "Representation, Surface", "Coloring", etc. But I do not see the "geometry view" (what is it?) and "Mesh Part" (as what you described).

What should I do at this point? Please be patient with me. I think I need some more experienced OF user to lead me through the "baby-steps". Thanks.

fengjq



Thanks for the advice. I'll give it a try and let you know how it works.[/QUOTE]

[ansubru]

Hi fengJq!!

Well you got the first bit right. You are able to visualize the domain right that is a foot forward in the right direction. Before i proceed any further, from what you have posted, I can gather that you are able to implement the solver without any errors and there are some computational results being produced as well?? If this is the case then read on...

Now, when you opened the vtk files, did you get something like this (look at the screenshot I have attached with this thread).

If your foamToVTK has been successful (which i think is the case) you should get a sub-folder as shown in the screenshot (of course with the relevant mesh parts). Along with the mesh parts named sub-folders, you would also get a folder labelled 'lagrangian'... this folder contains all the requisite particle specific data that has been generated.

Now in paraview, enter this sub-folder and open kinematiccloud.vtk (which is located inside the subfolder within lagrangian). Once you select this you should get something like this (look at relevant screenshot). click on apply.

Now, having highlighted kinematic cloud, generate the glyph (look at screenshot again)..

Set up the properties of the glyph (use the screenshot again as a reference.. remember you can alter the settings yourself to see what each value in the glyph means)..

You must be able to visualize the glyph (remember to forward the time step from 0 to 1 (look at screen shot carefully)..

If you have followed these steps carefully, I guess you must be able to visualize the hopper case..

Hope this clarifies your doubts

Best regards

ansubru

PS .. the screenshots are for a case that I have built (where I use the DPMFoam solver.. which is quite similar to the hopper case.. don't get confused by it.. its only a representative case..)

[fengjq]

Hi ansubru,

Thanks for offering more details with screenshots. I first tried with my "hopper" case following your steps and comparing your screenshots and still get what I had before. The ways I saw the difference here is that you said to open the *.vtk file in the "lagrangian" folder whereas I opened the *.vtk file in the VTK folder. But I tried with opening the *.vtk file in the "lagrangian" folder and didn't get what you had.

Thus, I went on to tried the "aachenBomb" case as in your example. Please tell me what I did wrong here.

1. cd .../lagrangian/sprayFoam/aachenBomb;
2. blockMesh, and run "sprayFoam" (which took hours on my PC);
3. after the solver finished, I saw data files for each time step were generated in the case directory aachenBomb;
4. foamToVTK, which generated a folder "VTK" where I saw bunch of aachenBomb_..vtk files (apparently at various time steps) and 2 subfolders "walls" (the "mesh parts"?) and "lagrangian";
5. paraview, Open > VTK > lagrangian > sprayCloud, double click on the "sprayCloud_..vtk" (which has a "+" symbol in front), then "Apply";
6. there is nothing displayed (not even the domain geometry) in the display window, unlike what you had in your screenshots;

Is there some other files I also need to open? For example, if I open the "aachenBomb_..vtk" right under the "VTK" folder, I get the (rectangular) geometry displayed in the paraview window.

By the way, what does this "aachenBomb" mean? what's the physical description of this case (which I think OpenFOAM can provide in a README to help the users)? I can understand the "cavity" case for incompressible flow as well as some other tutorial cases (just from the case title), but not "aachenBomb".

I noticed some differences in folder names between your screenshots and what I saw on my computer. It seems that you are running on Ubuntu? whereas I'm on OpenSUSE. Should that make any difference (even we both use OpenFOAM 2.3.0)? If so, I can set up a Ubuntu on my vbox and try it there.

In my OF 2.3.0 tutorial folder, I found "aachenBomb" under the "sprayFoam", not in the "DPMFoam". Does it matter?

Originally Posted by ansubru

Hi fengJq!!

Well you got the first bit right. You are able to visualize the domain right that is a foot forward in the right direction. Before i proceed any further, from what you have posted, I can gather that you are able to implement the solver without any errors and there are some computational results being produced as well?? If this is the case then read on...

Now, when you opened the vtk files, did you get something like this (look at the screenshot I have attached with this thread).

If your foamToVTK has been successful (which i think is the case) you should get a sub-folder as shown in the screenshot (of course with the relevant mesh parts). Along with the mesh parts named sub-folders, you would also get a folder labelled 'lagrangian'... this folder contains all the requisite particle specific data that has been generated.

Now in paraview, enter this sub-folder and open kinematiccloud.vtk (which is located inside the subfolder within lagrangian). Once you select this you should get something like this (look at relevant screenshot). click on apply.

Now, having highlighted kinematic cloud, generate the glyph (look at screenshot again)..

Set up the properties of the glyph (use the screenshot again as a reference.. remember you can alter the settings yourself to see what each value in the glyph means)..

You must be able to visualize the glyph (remember to forward the time step from 0 to 1 (look at screen shot carefully)..

If you have followed these steps carefully, I guess you must be able to visualize the hopper case..

Hope this clarifies your doubts

Best regards

ansubru

PS .. the screenshots are for a case that I have built (where I use the DPMFoam solver.. which is quite similar to the hopper case.. don't get confused by it.. its only a representative case..)

[ansubru]

Hi There

Originally Posted by fengjq

Is there some other files I also need to open? For example, if I open the "aachenBomb_..vtk" right under the "VTK" folder, I get the (rectangular) geometry displayed in the paraview window.

Yes, you need to open the .vtk files within each mesh part name (for eg the "walls" sub-folder etc).. You must then try to overlay different scenes. Lets take is step by step -

First, open the .vtk files for each mesh part. and finally the "aachenBomb_..vtk". Now use a representation type for each part, i use "outline" form as the domain is transparent, or the best way to visualize is.. open only "aachenBomb_..vtk" and use "volume" representation..

Once you can visualize your domain, open the .vtk files in "lagrangian" and follow the steps in the previous post. i guess that should be about it.

what does this "aachenBomb" mean? what's the physical description of this case

Read this...

http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2008/PerCarlsson/PC_Tutorial_dieselFoam_peered_NL_HN.pdf

It seems that you are running on Ubuntu? whereas I'm on OpenSUSE. Should that make any difference (even we both use OpenFOAM 2.3.0)? If so, I can set up a Ubuntu on my vbox and try it there.

Well, from my knowledge, I guess it should not be a major difference. I have never used OpenSUSE so I cannot explicit tell. Nevertheless, i believe the program should behave in the same way for OpenSUSE..

In my OF 2.3.0 tutorial folder, I found "aachenBomb" under the "sprayFoam", not in the "DPMFoam". Does it matter?

aachen bomb is a sprayFoam tutorial, as the solver has been developed for spray modelling of diesel injection. DPMFoam and srayFoam are two different solvers. The DPMFoam solver cannot account for bubble break-up or any-other phenomena specific to spray applications. FYI the tutorial for DPMFoam is the Goldschmidt fluidized bed case.

Cheers!!

ansubru

[fengjq]

Hi ansubru,

Thanks for the quick reply. This is great! Now I can see lots of stuff that I could not before, due to the missing steps with details settings (which is probably different with the Lagrangian cases from those of incompressible and compressible flow cases where we simply use parafoam to get everything loaded).

However, I still do not get what you showed in your screenshots. I guess now my problem is to figure out what kind of the "representations" to use for "sprayCloud" and "Glyphs". My displayed geometry is a rectangular box whereas yours (in your screenshots) seems different. You seemed to have particles displayed as (blue) spheres; but I don't know how to get that (I guess there must be some kind of settings for the particle size).

I couldn't get that webpage on "aachenBomb" to open. So I still don't understand the solver procedures for this case. I guess the solver must try to solve both the CFD problem for the flow field (+ thermodynamics) and the Lagrangian particle motions. With paraview we have to open several *.vtk files. I expect one of those should allow me to view the flow field as well as pressure, temperature, etc. and one to allow me to view particles and particle traces. I saw you seemed to have a display of the "U2" field in one of your screenshots. Would you please offer me further advices?

Thank you very much,

fengjq

Originally Posted by ansubru

Hi There



Yes, you need to open the .vtk files within each mesh part name (for eg the "walls" sub-folder etc).. You must then try to overlay different scenes. Lets take is step by step -

First, open the .vtk files for each mesh part. and finally the "aachenBomb_..vtk". Now use a representation type for each part, i use "outline" form as the domain is transparent, or the best way to visualize is.. open only "aachenBomb_..vtk" and use "volume" representation..

Once you can visualize your domain, open the .vtk files in "lagrangian" and follow the steps in the previous post. i guess that should be about it.



Read this...

http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2008/PerCarlsson/PC_Tutorial_dieselFoam_peered_NL_HN.pdf

Well, from my knowledge, I guess it should not be a major difference. I have never used OpenSUSE so I cannot explicit tell. Nevertheless, i believe the program should behave in the same way for OpenSUSE..



aachen bomb is a sprayFoam tutorial, as the solver has been developed for spray modelling of diesel injection. DPMFoam and srayFoam are two different solvers. The DPMFoam solver cannot account for bubble break-up or any-other phenomena specific to spray applications. FYI the tutorial for DPMFoam is the Goldschmidt fluidized bed case.

Cheers!!

ansubru

[ansubru]

Hi!!

Try this for the aachenbomb tutorial..

https://www.google.dk/url?sa=t&rct=j...63934634,d.bGQ

"representations" to use for "sprayCloud" and "Glyphs"

You just need to find the suitable representation for the glyph. It is the glyphs that contain the necessary tool to visualize temporal changes (changes in velo, temp etc over time). I use "sphere" to visualize particle (pls look at the screenshot in the earlier carefully.. all settings you need are given there.. just play around with them a bit). Use "arrow" representation to get particle velocity vectors.

You seemed to have particles displayed as (blue) spheres

Use "color by" option .. present in the toolbar on top (near the icon of the color scale). You can color by using any particle property while using a glyph on the kinematic cloud.

I expect one of those should allow me to view the flow field as well as pressure, temperature, etc.

To view these, you need to plot glyphs on the main "aachen....vtk". Remember to visualize any feature of the continuum, you need to perform your overlays on the main "aachen...vtk" file.

Cheers!!

ansubru

[fengjq]

Yeah, now I have the spheres displayed. Thanks a lot, ansubru. It appears like a dense plume of fuel droplets is injected into the (combustion?) chamber. After a while (around time step ~30), the droplets disappear (which I guess to evaporate by burning?). Then the rest time steps seem boring (nothing is showing up, probably depending on what I chose for display).

With the particle data files loaded, I guess I can now try the "particle path" filter. It seems that I can only have "particle path" with the "aachenBomb" data. Any advice on that?

Originally Posted by ansubru

Hi!!

Try this for the aachenbomb tutorial..

https://www.google.se/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CCwQFjAA&url=http%3A%2F%2Fwww.tfd.chalmers.se%2F~hani%2Fkurser%2FOS_CFD_2008%2FPerCarlsson%2FPC_Tutorial_dieselFoam_peered_NL_HN.pdf&ei=k_M6U_CvIKql4ATVjYCYBA&usg=AFQjCNGxgu9fJ9MZTlryZCt5dFzABpy45Q&sig2=eOxWucjIDxJ6Z1oOshJeYA&bvm=bv.63934634,d.bGE

You just need to find the suitable representation for the glyph. It is the glyphs that contain the necessary tool to visualize temporal changes (changes in velo, temp etc over time). I use "sphere" to visualize particle (pls look at the screenshot in the earlier carefully.. all settings you need are given there.. just play around with them a bit). Use "arrow" representation to get particle velocity vectors.



Use "color by" option .. present in the toolbar on top (near the icon of the color scale). You can color by using any particle property while using a glyph on the kinematic cloud.



To view these, you need to plot glyphs on the main "aachen....vtk". Remember to visualize any feature of the continuum, you need to perform your overlays on the main "aachen...vtk" file.

Cheers!!

ansubru

[ansubru]

Hi Again!!

I have personally not implemented 'particle tracks'.. however, there are loads of threads in this forum for this.. for instance.. i guess this one might help..

http://www.cfd-online.com/Forums/ope...-paraview.html

ansubru

[fengjq]

Thank you, ansubru. You've been a great help to me.

Originally Posted by ansubru

Hi Again!!

I have personally not implemented 'particle tracks'.. however, there are loads of threads in this forum for this.. for instance.. i guess this one might help..

http://www.cfd-online.com/Forums/ope...-paraview.html

ansubru

[pnovo]

Hi, I was directed into this thread with the very same problem and for my case I found the solution in this other thread.
So the 0 file in the lagrangian folder converted by foamToVTK should be deleted in order to be able to load the particles into paraview.
Cheers