Learning Forum

[ar097dr]

Hi everyone, I’m working on a 3D simulation of a wave rotor in ANSYS Fluent Student Edition. To simplify the computation, I’m taking advantage of symmetry and simulating only half of the rotor. I’ve attached screenshots showing the geometry and mesh for clarity.    Problem: When I run the simulation using the pressure-based solver, I encounter a Floating Point Exception error. If I switch to the density-based solver, I get a different error: Node 1 Fatal signal raised sig = Segmentation fault
 b3f1c3a0 CX_Primitive_Error
 ffd881f0 seh_filter_exe
 b3fc24d0 logical_right_shift
 e41be720 _C_specific_handler
 23e3dc0 _chkstk
 229e8c0 RtlWow64GetCurrentCpuArea
 23e3770 KiUserExceptionDispatcher
 b1c7c530 Partial_Slip_Thermal_Coef
 b3746890 Multi_Stage
 b3746200 Solve_Coupled
 b16beb80 alloc_mp_drag_coeff
 b0df6c40 Models_Send_update_solve
 b15ede90 solver_update
 b15ea0f0 Flow_Iterate
 b0cb8b00 CX_Disable_Interrupts
 b3fac470 eval
 b0dd0150 PRF_Command_Start
 b0dd2880 PRF_Node_repl
 b0dde690 Delay_Error
 b3fc24d0 logical_right_shift
 98e8c0 BaseThreadInitThunk
 235fba0 RtlUserThreadStart

 

What could be causing these errors in both solvers? Are there specific adjustments I should make for a half-geometry simulation using symmetry? How should I proceed to ensure the simulation runs without errors? Any guidance or suggestions would be greatly appreciated! Thanks in advance for your help!    

[Rob]

Where is the fluid going? I'm not sure what you're trying to model, but if it's a rotor where's the stationary bit, and how is the flow symmetrical? 

[ar097dr]

So basically the inlets and the outlets are located in the two stationary plates as you can see in the screenshot below. The other one is the rotating rotor with his channels.

 

For the symmetry, I'm talking about the fact that my geometry is cut in half because it is periodic and so around the circumference the same condition apply twice. Do you know why i keep getting those errors? 

 

[Rob]

So, you've got two sets of rotational periodic boundaries and two pairs of non-conformal interfaces? 

[ar097dr]

Basically yes 

[Rob]

Are you looking at temperature too? 

[ar097dr]

Yes the periodic boundaries conditions are in terms of pressure and temperature. 

 

[Rob]

Turn off rotation. What happens? If that fails turn off periodic, then try. Something isn't set up correctly, but the best way to figure out what that is is to turn things off until it works. 

[ar097dr]

So I turned off rotation and periodic leaving on only the boundary condition, then launched the simulation. If pressure based this is the error that pop up:
"

Updating solution at time level N...

done.

 

iter continuity x-velocity y-velocity z-velocity energy time/iter

1 1.0000e+00 1.0538e-02 8.4763e-03 2.2343e-02 6.7692e-04 0:00:15 19

 

Reversed flow on 203 faces (53.9% area) of pressure-inlet 27.

 

Reversed flow on 348 faces (99.8% area) of pressure-inlet 28.

 

Reversed flow on 36 faces (9.7% area) of pressure-outlet 29.

 

Reversed flow on 175 faces (48.9% area) of pressure-outlet 30.

Stabilizing pressure correction to enhance linear solver robustness.

Stabilizing pressure correction using GMRES to enhance linear solver robustness.

 

Divergence detected in AMG solver: pressure correction

Error at host: floating point exception

"

If density based this is the error:

Node 0 Fatal signal raised sig = Segmentation fault
 791bc3a0 CX_Primitive_Error
 c89481f0 seh_filter_exe
 792624d0 logical_right_shift
 adaae720 _C_specific_handler
 cad83dc0 _chkstk
 cac3e8c0 RtlWow64GetCurrentCpuArea
 cad83770 KiUserExceptionDispatcher

Any idea on how to fix it? 

[Rob]

Unless flow is doing about Mach 5 don't use Density based. There are exceptions, but I very much doubt you're modelling one of those!

Why pressure in and pressure out? 

[ar097dr]

Yes the flow is subsonic. Pressure in and out because my boundary condition impose certain pressure and temperature conditions on the inlet and outlet ports

[Rob]

Try fixing the flow and see how it behaves. Trouble shooting means eliminating potential problems and velocity in means you must see flow. 

[ar097dr]

Sorry I'm afraid I do not understand what are you asking me to do

[Rob]

Try using a velocity inlet. Initialise from that boundary and run a few iterations. What happens?

[ar097dr]

So I tried with velocity inlet, bust still breaks in the first time step after few iteration. I’ll copy the console report below:

Updating solution at time level N…

done.

 

iter continuity x-velocity y-velocity z-velocity energy time/iter

1 1.0000e+00 1.3271e-02 1.1309e-02 1.4667e-02 6.6454e-03 0:00:00 19

 

Reversed flow on 81 faces (20.9% area) of pressure-outlet 29.

 

Reversed flow on 167 faces (46.7% area) of pressure-outlet 30.

2 1.0000e+00 2.0676e-02 2.7264e-02 4.8245e-02 2.2009e-03 0:00:18 18

 

Reversed flow on 115 faces (30.4% area) of pressure-outlet 29.

 

Reversed flow on 18 faces (5.4% area) of pressure-outlet 30.

3 1.0000e+00 2.7187e-02 3.6026e-02 1.2450e-01 8.4653e-03 0:00:14 17

 

Reversed flow on 32 faces (8.4% area) of pressure-outlet 29.

 

Reversed flow on 25 faces (7.1% area) of pressure-outlet 30.

4 1.0000e+00 1.2537e-01 1.3589e-01 2.0873e-01 6.4304e-03 0:00:10 16

 

Reversed flow on 19 faces (5.7% area) of pressure-outlet 29.

 

Reversed flow on 13 faces (3.5% area) of pressure-outlet 30.

Stabilizing enthalpy to enhance linear solver robustness.

5 1.0000e+00 7.1412e-02 3.1269e-03 1.1736e-03 2.9869e-01 0:00:08 15

 

Reversed flow on 56 faces (15.7% area) of pressure-outlet 29.

 

Reversed flow on 45 faces (13.4% area) of pressure-outlet 30.

Stabilizing x-momentum to enhance linear solver robustness.

Stabilizing y-momentum to enhance linear solver robustness.

Stabilizing z-momentum to enhance linear solver robustness.

Stabilizing enthalpy to enhance linear solver robustness.

 

absolute pressure limited to 5.000000e+10 in 18 cells on zone 17

 

absolute pressure limited to 5.000000e+10 in 1 cells on zone 18

6 6.3736e+07 3.8551e-02 2.4861e-05 6.7488e-07 2.3187e-01 0:00:09 14

 

Reversed flow on 190 faces (51.9% area) of pressure-outlet 29.

 

Reversed flow on 52 faces (14.8% area) of pressure-outlet 30.

Stabilizing x-momentum to enhance linear solver robustness.

Stabilizing y-momentum to enhance linear solver robustness.

Stabilizing z-momentum to enhance linear solver robustness.

Stabilizing enthalpy to enhance linear solver robustness.

Stabilizing enthalpy using GMRES to enhance linear solver robustness.

 

Divergence detected in AMG solver: enthalpy

absolute pressure limited to 5.000000e+10 in 20 cells on zone 17

 

absolute pressure limited to 5.000000e+10 in 6 cells on zone 18

 

Error at host: floating point exception

 

[ar097dr]

P.S. It breaks even if I just set test value of velocity at SL condition

[Rob]

OK, so what do the velocity contours look like at iteration 2 and 3? Trouble shooting means more than just reading the console output. 

[ar097dr]

The point is that Fluent does not generate the animation to show the velocity contour. I tried using the Results section in Ansys, but I'm not very good at it. Do you have any tips?

[Rob]

Use Fluent. Create an isosuface (Mesh > Radial Coordinate I think) and plot on that. Just initialise and display run an iteration & repeat. If you're using Workbench, don't, and run Fluent in standalone mode.