OpenFOAM 1.6 and 1.7 with interFoam, groovyBC give different strange results
 

Hi all,

I have recently updated to OpenFOAM 1.7.0, after not having done simulations in Openfoam for a longer time.

I have a problem with a test case of a wave tank using interFoam and groovyBC on the inlet (left side) to generate waves in the flume. On 1.6 (or maybe it was 1.5), everything went fine, waves are running into the channel (left picture below). On 1.7.0, the same case with corrected p > p_rgh, gamma > alpha1 gives strange results and of course aborts after some time with a 'floating point exception'. Waves might be generated directly on the left inlet boundary, but the water is running up the front wall and down the back wall.

For net generation, Salome was used. The case files can be found at http://130.75.108.10/~material/pdf_downloads/20_wellentank_fine_complete.tar.gz

Pictures are attached. Left: OpenFoam 1.7, right, OpenFoam 1.6, for a similar time step.


Have you got any ideas of whats wrong here?

Arne

Greetings Arne,

I'm not familiar with groovyBC (yet) but I've got a feeling that it might be an issue with the compiler!

What gcc version(s) did you use to build those OpenFOAM versions and respective groovyBCs?

If you are wandering why I ask... this is the reason: http://www.cfd-online.com/Forums/ope...gcc-4-5-x.html

Best regards,
Bruno

Hi Bruno,

thanks for your hint. I'm running OpenFOAM on a 64bit Ubuntu 10.04.

A gcc -v gives me: gcc version 4.4.3 (Ubuntu 4.4.3-4ubuntu5). For the OpenFOAM 'installation', I did it the way described in http://www.openfoam.com/download/ubuntu.php.

Btw: Could anywone who is using groovyBC (1.6) and OpenFoam 1.7 test my case?

Arne

No offense. But have you checked the direction of gravity?

Check groovyBC
 

The direction of gravity may by wrong, check the orientation of your mesh and correct (or change the g-direction in constant/g). Check also the GroovyBC in 0/U, 0/alpha1. It also depends on the axes orientation(

valueExpression "(pos().y<=A*cos(-w*ti.......
.
.
variables "l=10;A=0.5;g=vector(0,-9.81,0);k=2*.....
)

Thanks mks,

it was the orientation of 'g' in the file. I shouldn't copy files from a tutorial case without looking if they fit for my case...

You know, always this problem with the wood and the trees...


Another question: What kind out boundary condition would you suggest for the outlet of the flume? The waves shall 'run out' of the flume without influencing the next incoming waves - so some kind of a spongle layer. Is there already one implemented?

Arne

well... if there is something ready to use, i have not found it... it is still an issue for me how to let the waves go "through" the wall. I just made the flume tank long enough:-)

Hey all,

I am trying to simulate a wave tank in OpenFoam 1.7.1.
I am able to get groovyBC working on my system. I followed the example of 2D stokes wave available on the WiKi pages. I really loved this boundary condition as it is relatively easier to understand and implement.
Problem is the reflection from the outlet boundary. Currently I am using zeroGradient boundary condition on the outlet boundary.
I was wondering if anybody came up with a solution to transmit the waves.
Has anybody tried the waveTransmissive boundary condition at the outlet boundary? Is it pertinent to this situation? If yes, I would like to know how to implement it.

I also read that the inclusion of sponge layer near the outlet boundary may solve the problem. Is anybody familiar with it?

Thanks a lot in advance.

Abhi