[Andrew2]

Hello,

In Fluent, is there guidelines or additional information for running simulations with sliding mesh besides the basic theory found in the theory guide? Is the sliding mesh operation serial or parallel? Is there any way to view the time it takes for the mesh to be rotated? And for periodic repeats, do the two sides of the interface have to have the same angle of periodicity?

Thank you,
Andrew

[CeesH]

About the time for rotation: The number of timesteps depends on your setting of course. My advice would be to set an integer number of timesteps per rotation; I've once heard 30 as a ballpark number, but more likely this is mesh-dependent (I typically set it such that it moves 1 gridcell per timestep, so if I have 60 divisions tangentially, 60 timesteps for a full rotation).

Serial/parallel: whatever you wish, you can do both. For stability, I prefer setting up in serial, but you can run in parallel.

Periodic boundary conditions: I've never tried with any setup where that was not the case; but I think that indeed the rotating and stationary frame will need to cover the same angular range. Also, don't forget to set the interface between rotating and stationary frame to 'periodic repeats'

[Andrew2]

Hi,

My question was pointed more to the operation of the sliding mesh. I wanted to know how long it takes for fluent to rotate the mesh. Fluent tells you how long it takes to do a time iteration, but there is a long pause at the beginning of the time iteration. I assume Fluent is rotating the mesh in this pause, but I can't seem to find an option to view how long this operation take. I would like to know so I can try to optimize my setup, i.e number of processors.

For the serial/parallel question, I wanted to know whether the operation to rotate the mesh is parallelized, or is serial? From experience, the solver has choice of being serial or parallel, but what about the operation to conduct the sliding mesh?

I know ideally the angle of periodicity for the rotor and stator should be equal, but for most real application, the number of rotors and stator don't match. If the angle differ by 2 or 3 degrees, will greatly affect the solution?

Thank you,
Andrew

[CeesH]

Ah, I see; can't help you on that one. But I assume the rotation is parallel, at least, the cell partition on each core will need to be rotated.

Related to the rotor/stator problem, I think that depends a lot on your setup. In my case, I simulate an impeller with 6 blades, in a tank with 4 baffles. Running a 60 degree periodic domain (with essentially 6 blades and 6 baffles) has hardly any influence on the flow, however, since rotor-stator interaction is weak. In case this is strong, the number may be more important.

[Far]

Originally Posted by Andrew2

Hi,

I know ideally the angle of periodicity for the rotor and stator should be equal, but for most real application, the number of rotors and stator don't match. If the angle differ by 2 or 3 degrees, will greatly affect the solution?

Thank you,
Andrew

Then you need to scale your geometry to make the pitch equal on both sides.

Please consider CFX for this case. search for transient blade row (TBR) models in CFX help.