[DaveR]

Hi all,

I have been trying to validate my OpenFOAM methodology by modelling a 3-bladed water turbine and validating it against literature as found here.
Ultimately I want to match recorded torque values over the blades through refinement of the boundary layer region, however at the moment I am just trying to get a simulation running.

I have spent a long time trying to obtain a semi-reasonable mesh to solve using pimpleDynamicMeshFoam. I feel like my current mesh is good enough to get atleast something out. I've attached a screenshot and welcome your thoughts on this.

I am solving using the k-omegaSST turbulence model, however I am little confused about the calculation of input parameters in the 0 folder.
I have calculated the input variables nut, k, omega using the turbulence guidelines found here, and verified them through the calculator here. Velocity is taken from the aforementioned literature, p is set as zeroGradient.

My simulation currently fails through the Courant number exploding, resulting in a floating point exception early in the simulation. Given that the Courant number is governed by u*t/x, in which freestream velocity and time step are stated in literature - this would initially lead me to believe the mesh is still the issue.

However, Para View doesn't seem to offer any hints as to a location in the mesh that could be failing, so perhaps this is likely to result from my boundary conditions/input parameters?

Is anyone able to shed any light on my thought processes on setting up the simulation correctly using turbulence models, or perhaps where else I may be going wrong?

I've attached my case at the following google drive link if anyone could be so kind to take a look and offer any advice on getting this running and heading towards validation - it's been troubling me for some time now!

https://drive.google.com/open?id=0B3...UJxc1BpTS1XZDQ

Best regards,

Dave

[DaveR]

For ease of reference my simulation criteria have been calculated as follows:

Literature values:

Turbulence model: kOmegaSST
U = 2.8 m/s
Tip speed ratio = 2
Turbulent Intensity, I = 2%
Chord Length = 0.032 m
Time step = 1 rot deg/time step = 2.707e-04 s

Calculated values:

p = constant
Turbulent Viscosity, nut = 2.4e-03
Turbulent Energy, k = 4.704e-03
Specific Dissipation Rate, omega = 1.9595
Turbulent Length Scale = ~5% of inlet height = 0.035
Angular velocity, omega = 64 rad/s


Anything particularly obvious that could be wrong?

[tareqkh]

Hello Dave,

I have done similar simulations before and found some good case files that they might be helpful in terms of the setup etc. One of the cases works fine on OpenFoam v 2.3.2. The other case file requires some work to get it to work. Anyway, the file are uploaded in mediafire;

http://www.mediafire.com/file/4mia94y679r645d/OpenFOAM-2D-VAWT-master.zip

http://www.mediafire.com/file/n8sjaavaapcqvay/Naca-TSR1.0.zip

Please let me know if you find them helpful.

Best,

[DaveR]

Hi Khamlaj,

Thanks very much for those, I'll look over them today and report back any findings!

All the best,

Dave

[DaveR]

Hi again Khamlaj,

The second mesh seems to be faulty as it doesn't load in ParaView, detailing the error:

--> FOAM FATAL ERROR:
Symmetry plane 'symmetryPatches' is not planar.
At local face at (-8.82 -9 0.005) the normal (0 -1 0) differs from the average normal (0 0 0) by 1
Either split the patch into planar parts or use the symmetry patch type

From function virtual void Foam::symmetryPlanePolyPatch::calcGeometry(Foam::P streamBuffers&)
in file meshes/polyMesh/polyPatches/constraint/symmetryPlane/symmetryPlanePolyPatch.C at line 64.

FOAM exiting

The first mesh seems to solve well, although not at the time step stated in the control dictionary. I've modified this case to run on my own polyMesh files and the simulation quickly fails with the Courant number exploding.

Do you think this is then indicative of a mesh problem, rather than the initial conditions listed in my 0 directory?

I have also wondered if simulation failure is related to the fvSchemes and fvSolutions dictionary.. What are your thoughts on this? I don't feel I understand these dictionaries very well - perhaps this could be the root of the problem?

Thanks again for your time,

Dave