[sushancfdonline]

Hi,

We are using Ansys 2020 R2 product for CFD analysis. We have used CFX for RANS based simulation of a counter rotating axial compressor stage. Both are rotors and rotating in opposite direction (no stator). We are happy to share that CFD results show good matching with the experimental data. For the further analysis, we are trying to simulate a full annulus unsteady run for the same stage. My flow regime is subsonic only.


Using ANSYS CFX, I tried to run a transient case of full annulus contra rotating axial fan using the Transient Rotor Stator model at the interface of both rotors. In ANALYSIS TYPE I selected the TRANSIENT BLADE ROW option with Time period 0.02 seconds and time steps 0.0001 seconds. But I failed in getting a continuity in contours at the interface of both rotors. For clarity I here attached the image I got. My doubt is, within TRANSIENT BLADE ROW, different options are there - Profile transformation, Time transformation and Fourier transformation. I have tried TRANSIENT BLADE ROW with TIME TRANSFORMATION but the entropy contours are not matching at the interface.

Earlier I tried the TRANSIENT option with TRANSIENT ROTOR STATOR model at the interface, but the same image I got. I need clarification on why the entropy contours which represents the wake regions of the first rotor seems scattered when it enters into the second rotor domain. I am using the full annulus domain (not single passage). Both the discontinuous images I got are attached here. Please kindly help me in solving this problem.

Attachments: https://im.ge/i/TCskXm
https://im.ge/i/TCsVZf

[Opaque]

Would you mind adding the mesh lines on both sides of the sliding interface?

It seems as if the wake is being diffused when crossing the interface. Is the mesh downstream coarser or lower quality than the mesh upstream?

[sushancfdonline]

Thanks for the response Opaque. Both are fine meshes only, but total number of elements at the interface would be different since the number of blade in rotor1 (9 blades) and rotor2 (7 blades) are different. Is this mismatch in the number of mesh elements at the interface creates such problems?

[Opaque]

If the mesh quality is good on both sides, this should not happen.

I would focus on the transient, or transient blade row + profile transformation setups only until figuring out what is off with your case. Both solutions should be identical assuming the setups are correct.

Would you mind sharing only the CCL section for the sliding domain interface?

[sushancfdonline]

Originally Posted by Opaque

If the mesh quality is good on both sides, this should not happen.

I would focus on the transient, or transient blade row + profile transformation setups only until figuring out what is off with your case. Both solutions should be identical assuming the setups are correct.

Would you mind sharing only the CCL section for the sliding domain interface?

FLOW: Flow Analysis 1
&replace DOMAIN INTERFACE: R2 to R1
Boundary List1 = R2 to R1 Side 1
Boundary List2 = R2 to R1 Side 2
Filter Domain List1 = R2
Filter Domain List2 = R1
Interface Region List1 = INBlock INFLOW 2,INBlock INFLOW 2 2,INBlock INFLOW 2 3,INBlock INFLOW 2 4,INBlock INFLOW 2 5,INBlock INFLOW 2 6,INBlock INFLOW 2 7
Interface Region List2 = OUTBlock OUTFLOW,OUTBlock OUTFLOW 10,OUTBlock OUTFLOW 3,OUTBlock OUTFLOW 4,OUTBlock OUTFLOW 5,OUTBlock OUTFLOW 6,OUTBlock OUTFLOW 7,OUTBlock OUTFLOW 8,OUTBlock OUTFLOW 9
Interface Type = Fluid Fluid
INTERFACE MODELS:
Option = General Connection
FRAME CHANGE:
Option = Transient Rotor Stator
END
MASS AND MOMENTUM:
Option = Conservative Interface Flux
MOMENTUM INTERFACE MODEL:
Option = None
END
END
PITCH CHANGE:
Option = Automatic
END
END
MESH CONNECTION:
Option = GGI
END
END
END