code_saturne User's Forum

Dear developers,

Hello. I would like to do LES on a turbomachinery cascade, demonstrated in the figure below. There will be two domains: the front one is a stationary domain (including the stationary vane), the rear one is a rotating domain (including a rotating blade). For the rotating domain, I would like to solve it in a rotating reference frame, by using "Turbomachinery module - Frozen Rotor" option.

The model is a sector, instead of the whole circle, by setting "rotational periodic" boundary conditions. The stationary domain equals 1/33 of the full circle, while the rotating domain equals 1/44. Therefore, two domains are not perfectly matched at the interface.

Could you please suggest if this is possible for Code_Saturne LES? Are rotor-stator interface and rotational periodic compatible?

Thank you very much in advance for any help!

Best regards,
Ruonan

Hello,

Transient rotor staror simulations are compatible with periodicity only if the periodic interface is not adjacent to the rotor stator interface, so would not work in your case.

A frozen rotor approach does not have this constraint but I am not sure whether it makes sense with LES.

Also, the default second order scheme used by LES (Crank-Nicholson) requires values at faces if I remember correctly, and is thus not compatible with rotor stator either (we would need to interpolate values which are currently rebuilt instead). The double backwards Euler scheme would be an alternative, but is less precise with small time steps (and better with larger time steps).

Best regards,

Yvan