[Stroopwafelandcoffee]

I'm currently in the process of setting up a rotating reference frame simulation of an enclosed propeller. The mesh is generated in Pointwise, I've tried to get the quality as good as possible and am still talking to Pointwise to further improve it. I've also tried asking ANSYS about it, but we're not really getting anywhere. So I was hoping CFD-Online forum goers might be able to help out here. I've also added a link to the case so people can examine it themselves.

Settings:

General:

Steady

Models:

- Energy: On
- Viscous: SST k-omega

Materials:
- Air ideal-gas (all constant values for mat. prop.)

Cell zone conditions (1 zone):
- Frame Motion
- Relative to cell zone: absolute
- Zone motion function: none
- Rotation origin: 0 0 0
- Rotation axis: 0 0 1
- Rotational velocity: 147 rad/s
- Translational velocity: 0 0 0 m/s

Boundary conditions:
Inlet:
- Pressure Inlet/Outlet
- All standard settings

Wall moving with domain:
- moving wall
- relative to cell zone: 0
- rotational
- Origin: 0 0 0
- Axis: 0 0 1
- No Slip
- 0 heat flux and heat gen

Wall standing still:
- moving wall
- absolute: 0
- rotational
- Origin: 0 0 0
- Axis: 0 0 1
- No Slip
- 0 heat flux and heat gen

Inner radius 0 (point)

Outer boundary:
- stationary free slip wall

Two periodic domains:
- Rotational: 120 degrees

Mesh report:
Poor Mesh Element Statistics:
0 faces with too small area
0 faces adjacent to negative volume cells
0 faces adjacent to bad quality cells
0 left handed faces on mesh interfaces

Orthogonal Quality:
0.1% of cells below 0.1416 (this is what I'm talking to pointwise about)

Error:
Error at node 7: Specific_Heat_by_T_Integral: Invalid lower bound, T0 <= 0
Error at node 13: Specific_Heat_by_T_Integral: Invalid lower bound, T0 <= 0
Error at node 10: Specific_Heat_by_T_Integral: Invalid lower bound, T0 <= 0
Error at node 2: Specific_Heat_by_T_Integral: Invalid lower bound, T0 <= 0

I'm also curious how I can find out which nodes these are. Can I zoom in on the location of a node by ID? I have not yet found a way to do so.

Case:
https://drive.google.com/open?id=1Ct...gNrsQrMHWd2T3i

Please let me know what you think. Any help would be greatly appreciated.

[MKuhn]

I don't know Pointwise. But it seems that the mesh is not suitable for Fluent, when even "Standard Initialization" fails. As Fluent crashs after the solution initialization, you will be not be able to locate the error nodes.

With your case-file, I get warnings like:

Note: zone-surface: cannot create surface from sliding interface zone. Creating empty surface.

And trying to convert to Polyhedra:

Cell Equivolume Skewness too high.

Info: Used modified centroid in 599 cell(s) to prevent left-handed faces.

And after initialization, like you

Error at Node 1: Specific_Heat_by_T_Integral: invalid lower bound, T0 <= 0.
Error at Node 0: Specific_Heat_by_T_Integral: invalid lower bound, T0 <= 0.
The fluent process could not be started.

Ask Pointwise again, what is going wrong with their bloody software.

[Stroopwafelandcoffee]

Originally Posted by MKuhn

I don't know Pointwise. But it seems that the mesh is not suitable for Fluent, when even "Standard Initialization" fails. As Fluent crashs after the solution initialization, you will be not be able to locate the error nodes.

With your case-file, I get warnings like:

Note: zone-surface: cannot create surface from sliding interface zone. Creating empty surface.

And trying to convert to Polyhedra:

Cell Equivolume Skewness too high.

Info: Used modified centroid in 599 cell(s) to prevent left-handed faces.

And after initialization, like you

Error at Node 1: Specific_Heat_by_T_Integral: invalid lower bound, T0 <= 0.
Error at Node 0: Specific_Heat_by_T_Integral: invalid lower bound, T0 <= 0.
The fluent process could not be started.

Ask Pointwise again, what is going wrong with their bloody software.

Could you perhaps verify that at least my simulation setup is in order? I would like to be able to at least rule out any mistakes in that part of the problem if I can.

[MKuhn]

It seem that there some problems with your interface between perin and perout. If you delete the interface and switch it to normal wall, than the initialization works.

[Stroopwafelandcoffee]

Originally Posted by MKuhn

It seem that there some problems with your interface between perin and perout. If you delete the interface and switch it to normal wall, than the initialization works.

So here's how the interface was created:

A rotational periodic domain was created in Pointwise, to ensure point matched boundaries. The "Interface" BC was assigned to them and the case was exported to Fluent. Here the following TUI command was used to make both boundaries periodic:

/define/mesh-interfaces/make-periodic perout perin yes 120.0 yes yes periodic

I was thinking: Maybe I'm interpreting the "Interface" BC type wrongly.
I'm now trying the Symmetry condition

Alternatively, I have in the past also tried the following TUI command:

/mesh/modify-zones/make-periodic perout perin yes yes

But this also gave the same error. I have read the following thread that also speculated that the periodic domains caused this exact error:

Hybrid Initialization Error

It suggested using

grid/repair-improve/repair-periodic

But this does not help my case either, the error remains.

Is there a possibility of checking exactly what is wrong about the periodic interface?

[MKuhn]

Originally Posted by Stroopwafelandcoffee

Is there a possibility of checking exactly what is wrong about the periodic interface?

Sorry, I have no more idea.