[akrasemann]

Hi,

first of all a summary of my installed software:
Ubuntu 13.10
OpenFOAM 2.2.2
ParaView 3.12.0
enGrid 1.4.0

What I would like to achieve is a set up capable of assessing both heat transfer (conduction, convection and radiation) and flow in a rather complex structure. In best case a 3D steady state model as well as a 3D transient model will be achieved.

Starting from both the tutorial cases 'multiRegionHeater' for the transient and steady state version of the chtMultiRegion-solver, I'm in the process of adapting the set up to my own needs, but useing a plane wall (and neglecting radiation) in order to link results to analytical solutions in a first step. By doing that I face the issue that my temperatures do not converge and causes the solver to stop with the following error:

[2] --> FOAM FATAL ERROR:
[2] Maximum number of iterations exceeded
[2]
[2]     From function thermo<Thermo, Type>::T(scalar f, scalar T0, scalar (thermo<Thermo, Type>::*F)(const scalar) const, scalar (thermo<Thermo, Type>::*dFdT)(const scalar) const, scalar (thermo<Thermo, Type>::*limit)(const scalar) const) const
[3]
[3]
[3] --> FOAM FATAL ERROR:
[3] Maximum number of iterations exceeded
[3]
[3]     From function thermo<Thermo, Type>::T(scalar f, scalar T0, scalar (thermo<Thermo, Type>::*F)(const scalar) const, scalar (thermo<Thermo, Type>::*dFdT)(const scalar) const, scalar (thermo<Thermo, Type>::*limit)(const scalar) const) const
[3]     in file /home/opencfd/OpenFOAM/OpenFOAM-2.2.2/src/thermophysicalModels/specie/lnInclude/thermoI.H at line 76.
[3]
FOAM parallel run aborting

By using the search function within this forum I got several hints on solving this issue:

• Mesh quality
• Boundary and initial conditions
• Changing the number of iterations within the called function (not recommended)

In the following a description of what I was doing:
https://www.dropbox.com/s/x7b35dsaj6ao6r7/Geometry.pdf

As I set it up as a 3D case, although I am aware that for this problem a 2D case is sufficient, apart from the depicted patches (in red) there are <regionname>_front and <regionname>_back patches.

Regarding the mesh:
The mesh is generated in enGrid (1.4.0) and then exported to OpenFOAM using the OpenFOAM (polyhedral) export function. My mesh related questions were already answered by wyldckat: http://www.cfd-online.com/Forums/eng...gion-case.html
I am aware that a mesh generated by the blockMesh utility would suit that problem very well, but as I would like to go to a more complex structure, I try using the same process of setting up the case as I would do for future application. Due to the straightforward boundary layer refinement implemented in enGrid I would like to follow that path.
What is important to note is that by using the enGrid export function, the OpenFOAM set up does not follow the process of splitMeshRegions and changeDictionaryDict to account for multiple regions.

Regarding Boundary and initial conditions:

As I'm using a k-epsilon turbulence model, I had a look at http://www.cfd-online.com/Tools/turbulence.php, the referred explanations on the single entities that are used for calculating k and epsilon as well as the OpenFOAM user guide, which explains on these within section 2.1.8 (Lid driven cavity flow tutorial).
I ended up with the following equations:
k = 3/2*(sqrt(U_x^2 + U_y^2 + U_z^2)*I)^2, where U_i are the mean flow velocities and I is the turbulence intensity
epsilon = (C_mu^(0.75)*k^(1.5))/(l_t), where C_mu is a constant (0.09) and l_t is the turbulent length scale, which needs to be estimated.
For the plane wall case I use l_t = L_char * 0.4*delta, where L_char is a characteristic length (1.5m x-dimension) and delta the boundary layer thickness (4.91*L_char/sqrt(Re_x))

I then set up the case using the derived k and epsilon values, according to the geometry, fluid properties and individual flow. Interestingly I face the reported error in the laminar region (U=0.1, U=1), but a turbulent case (U=10) works fine:
https://www.dropbox.com/s/si9y50brwc...ady_U10.tar.gz
Increasing the velocity to U=50 causes again troubles.
None of the transient settings worked out:
https://www.dropbox.com/s/gj62ozuik5...ent_U10.tar.gz

To resolve that i had a closer look at http://openfoamwiki.net/index.php/Ge..._-_planeWall2D and the OF2.2 case. I ran the provided case (U0.1) without any troubles. Even U=1 without changing k and epsilon worked fine. Switching to a turbulent scenario (U=10) and applying the above mentioned equations to the scenario produced a solution.
By the way I adapted the tutorial for running it in parallel. Please check the ChangeLog:
https://www.dropbox.com/s/wwxv9y7zzi...arallel.tar.gz

I realized that the k and epsilon values for the laminar region differs enormously from mine, so I used these values for my own set up, but again the solver crashes due to divergent temperature solutions, as stated above.

So I really am confused about how to set the case. I also do not see big differences in the boundary and initial conditions after looking at them for two days. All fvSolution and fvSchemes are exactly the same.

Sorry for the lengthy explanation, but I'm happy to share my documentation with the community after I resolved these issues.

Greetings,

Andreas