[Trinxo123]

Hello,

This is my first post here, so thanks in advance for trying to help.

I am currently investigating the thrust and torque variables of an axial fan (or axial impeller) that works for ventilation. I have tried simpleFoam with the Moving Reference Frame option (positioned just in the volume enclosing the propeller) and found some mismatches with the tested values (experimentally). So now I am trying different approaches, Sliding Mesh is one. However, it is incredibly expensive so I decided to try periodicity as well.

Geometry: The fan itself is composed of an external structure to distribute the air towards the impeller, the impeller, the engine, and another external structure for the fan wake, in order as the flow sees them. See the picture attached. Please note that I cannot post real images given that I am not the owner of this geometry (my university is).

Case: The fan is rotating clockwise at 1480 rpm, with incompressible flow and a "hovering" state, so the air is calm and is the fan itself that creates suction to speed up the flow. I have done a GCI and believe the mesh is not the problem, I am using Fluent Meshing and fluent3DMeshToFoam (with polyhedron).

Current simulation: I have split the volume into 8 wedge parts (one per blade) and tried to simulate it using simpleFoam with MRF on the whole volume (see picture 2 for the simulation volume), RAS turbulence with K-w SST. It seems that the MRF with the entire domain being rotative crashes the boundary conditions and a lot of artefacts appear. So I switched to SRFSimpleFoam, where by default the whole domain is rotative. The artefacts when away, but still I am not getting physical results.

The simulating wedge (or pipe as I call it) is sufficiently long so at the inlet and outlet BC, I am using fixed pressure to 0 Pa. I think that my BCs are the problem, so I describe them below (also check the 3th picture).
Inlet: zeroGradient for Velocity and fixedValue for pressure, as is the fan that creates suction, I can't determine an inlet velocity.
Outlet: inletOutlet for Velocity and fixedValue for pressure
Pipe walls: Slip walls
Periodic (right and left wedge walls): Cyclic AMI
Fan walls and Engine: SRFVelocity BC at 0 velocity.
Rotative parts (impeller and rotor): noSlip walls

What happens is that the flow seems to accelerate to infinity and the fan is driven by the flow instead of being the fan moving the flow, as I get negative torque in the rotative walls. Despite I impose 0 pressure at the inlet and outlet, a sudden increase in the pressure happens after the inlet so there is a clear gradient in the pipe from high to low pressure, which is contrary to the normal duty of the fan, (see 4th picture).

I have tried MRF with only the impeller zone rotative, but it crashes the cyclic BC all the time, so SRFSimpleFoam seemed the logical choice. Please let me know if you have any suggestions, tips, or solutions for this case. Once again, thanks a lot. I am also attaching the files Vrel, p, k, omega, and nut of the 0 time. If you have any questions, please ask.

IMPORTANT: I am uploading all pictures and dictionaries to weTransfer as I am having problems uploading them here (webpage is crashing) weTransferLink: https://we.tl/t-ffBoZTyhot

Thanks,

Marc