Motorbike Tutorial and No velocity data
 

Hello Everyone,

I manage to generate the mesh and run a solution. I can plot the pressure contours on the surface of the bike but there doesnt seem to be any velocity data.

Because of this, if a plot the velocity contours on the bike surface it remains a "single color". Also, Im not able to generate streamlines (and this is what i think) because there is no velocity data to plot.

Thoughts on why its not solving for the velocity field?

Thank you

Hi,

If you are using the 1.6 version, then it's only natural that it won't work all that well. I've tried running it with different architectures and the residues would always go to smithereens at around 300s... Try the 1.6.x version, it works waaay better!

Best regards,
Bruno

Bruno,

I have the directory OpenFOAM.1.6.x in my OpenFoam directory.

Do I simply type "git pull" to update?

Also how how are you plotting the residuals?

Thank you,

Adam

Adam,
You can try it, when inside your OpenFOAM-1.6.x folder, but depends on how long ago you did the last update. The version I've used was from late October. I haven't visually checked the git tree for a while now (I use gitk), so I don't know if there were any recent changes that might have blown the case's cohesion.

With the version 1.6, I just looked at the output from simpleFoam, while trying to figure out why Paraview couldn't see anything at 500s. At first saw the dead numbers (INF or NAN) at 500s, then tracked upwards to when it started to go sour.

I don't usually plot the residuals. But by what I've read, PyFoam seems to be one of the best choices. Oh, and a recent thread has a few more tips on plotting residuals!

And you're welcome :)
Bruno

Bruno and to anyone else following this post,

I am new to Linux and OpenFoam so to get gnuplot I typed in the terminal

sudo app-get install gnuplot

Please correct me if this was not the best method.

Then use the script found here so you can view the residuals real time:

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

I hope this helps.

Thanks again Bruno

Also, the solution looks converged but when attempting to plot in paraview its seems that there is no velocity. Im loosing my hair.

Hi Adam,

This is a wild guess... but did you check the other time snapshots? You have one every 100s... and only the 0s one is dead in the water! So, I'll almost bet you haven't used the time controls on the top of the Paraview window ;)

If that's not it, try this at the motorBike folder:
And open one or more .VTK files that there exist now. The "-latestTime" will give you the 500s snapshot. There you must have wind speed!! Hell, the motorbike picture we have at blueCAPE's website (found here), wouldn't exist if the 500s snapshot didn't have wind speed!

If that still won't work, try downloading the prebuilt paraview available here, and use it to open the VTK files!

As for (minor glitch you have there):
It usually works for Debian based Linux distros, like Ubuntu is.

Best regards,
Bruno

Bruno,

I usually run paraFoam from the working directory (per the tutorials) and in paraview I do make sure I am at the appropriate time step. It's interesting that I can't use paraFoam on the motorbike tutorial.

As you recommended I ran vtk and opened the vtk file in paraview and it contains all the appropriate data so thank you once again! Any thoughs as to why this vtk file contains the data and why I wouldn't be able to view the same information by running paraFoam?

Also, thanks for that link to the pre-compiled version of OpenFoam for windows, very cool. Do you prefer windowns over a linux OS? Has anyone compared convergence speed for the same case on both OS's?

Hello Adam,

For the past two weeks or so, I've been reading various threads that report on paraFoam not working properly. They range from bad downloads to problems with recent versions of Qt and Linux (Ubuntu 9.10 and OpenSUSE 11.2), and some have problems with miss-installed graphic cards in Linux (OpenGL issues). Some are problems with Paraview's core, which isn't fully compatible with recent versions of Qt... because although they build properly, they don't work all that well.
So, you have 3 options:

1. rebuild Paraview and the OpenFOAM plugins with Qt 4.3.5 as intended... and might want to use a version of Linux from 6 months ago;
2. use foamToVTK, and wait for a Paraview more recent than 3.6.1 to come out;
3. try and use the "native" reader that doesn't require OpenFOAM at all, available here.

Also, if you have issues with bad rescaling on Paraview, or some other issues, this thread has a few pearls of wisdom ;) Which is in fact the thread dedicated to the third option.
If you go through the threads of the last few days, you might find what I've talked about... I'm just not posting the links here, because there are about 5 or 6 of them :(

You're welcome. I didn't know you would be interested in this. Check out the benchmark we made with motorBike, which compares the various runtimes. I honestly think that running OpenFOAM is best in Linux, but conditions aren't always perfect and in those situations, building for Windows is the only choice possible... you know the old saying: «the client is always right». At least, while virtualization isn't fully mainstream :cool: And yes, Linux is that good... it can run inside a virtual machine and still be faster than Windows in a real machine!
Oh and a few more useful links about that package can be found here.

Best regards,
Bruno

residuals
 

for residuals this is the command that i use

solver_name log & pyFoamPlotWatcher.py --silent log

this plots the residuals and also the continuity vs cummulative curves

i have a similar problem where paraview doesnt simply show me the output to one of the variables (that i have created by modyfing the solver)
how could i get it to display the output????

Have you tried converting the final time step solution directory to vtk as suggested in the post??

not as of yet .....paraview hasnt recognised the addition of a new variable ...everything is working just fine with P and U.....but then nothing for the new varible ...could it a be a compiling problem??

thanks alot stark i'll still too try converting it to vtk and see what happens....

solved
 

ok i figured it out....usually paraview doesnt always give the variable output ...one would have to select it.

So here is my next question:

I am now able to view the mesh via paraview however I can only view the biker when I display the wireframe.

I would like to view the biker as a surface but, when I attempt to do this, it shows the fluid domain thus hiding the biker.

Thoughts on how to view only the bike when post-processing?

Thank you everyone.

Use Filters->Alphabetical->Extract Block to extract all motorBike_* surfaces if you just wish to look at the biker itself. Use Filters->Alphabetical->Slice for taking slices through the domain.

Unfortunatly the extract block option is not available. Thank you for the quick response!

Flow Visualization
 

I have a quick question in regards to visualizing the flow in ParaView...

I pretty much copied and pasted the motorBike example and put in an aircraft fuselage in there, however, I can't seem to figure out how to visualize the flow in a manner as done here (motorBiker.png/from blueCAPE website). I have the streamtracer turned ON, but it seems to me that I'm missing something big.

Can anyone help guide me along?

Also, kind of on another topic; the motorBike example is analyzed until 500 s? but why when you open it up in ParaView, the time goes from 0 to 1?

Thanks in advance!:)
Aleksey

I solved my problem :)

I was running the case under parallel, and I forgot to run the "reconstructPar" command after simpleFoam analysis. After I ran the reconstruction, everything fell into place (the flow visualization and time).

Aleksey