[TurBoris]

Hello,

Since TBR Fourier Transform does not support the inlet and outlet boundary disturbance with flutter, I have set up a flutter case with full wheel URANS model. I directly imposed the vibration input which comes from the modal analysis to blade. When I run the case it warns me with negative volume mesh at first iteration and stops the solver (TBR FT run with uniform inlet and outlet BC with this vibration input runs without any problem). When I check the coordinate of the negative volume cell it generally stands near blade tip & shroud.

I have shared the contour of inlet Pt profile, mesh displacement profile on blades, blade BC detail and shroud BC detail as visual.

In URANS, CFX does not ask for nodal diamater, rather it asks for phase anglee. I have calculated it as 360*4/22= 65.4545 degree for ND=4

What would be a reason for this error ? Does doing flutter in URANS differ from TBR cases ?

Thank you.

[Opaque]

Nodal Diameter option is not available for the Transient Analysis, but it is for Transient Blade Row Analysis.

Since you are using a full-wheel/360 [deg] model, you can use the Transient Blade Row Analysis, just do not activate any of the transient pitch change models. That is, let it be the default shown.

[TurBoris]

It worked, thank you very much. I have just activated TBR time integration as seen in visual. I have just defined the time step size as in transient case. In addition to transient case it requires only the time period. I set it as the duration of one shaft revolution, although the vibration period is half of it (because my inlet profile is 360 degree). Is there any mathematical difference between doing full wheel transient and full wheel TBR (without any pitch change model) ?

[Opaque]

None as far as I know.

For time-periodic problems, it is easier to control the time integration details and some features are only available via the Transient Blade Row option.

[TurBoris]

Hi again,

When I activate export surface Engine Order output option with integration method of full revolution CFX warns me that the run duration must be at least 6.283185 s which corresponds to 1680 wheel revolution. This lasts a couple of months to run. ( In my setup 1 revolution = 1 period is 0.00374 s)

Interestingly 6.283185 is 2*pi. How can it be associated with multiples of pi ?

How can I handle this problem ?

Thank you for any help

[Opaque]

I would check the Number of Passages per 360/Component settings.

It definitely needs the time it takes the rotor to complete one revolution since it says "Full Revolution" for the integration method.

Now be careful confusing Period in the TBR panel, and period for 1 full revolution. Something seems to be off by 2*pi as you noticed.

[TurBoris]

Thank you for reply.

I have defined cosine profiles at inlet and outlet BC as disturbances. The inlet profile provides EO 5 excitation and outlet EO 19. Since these disturbances are periodic at 1 revolution I set the time period as duration of 1 revolution of the wheel. Is that a correct approach ?

I select full revolution time integration for engine order output because at the end of 1 wheel revolution the signal will complete its period.