[D. Lehnardt]

Hello everyone,

I am using Star CCM+ to model a helicopter rotor in hover flight conditions using a periodic domain of just one blade.

My boundary conditions are illustrated in the first image.

I have assigned rigid body motion to the entire domain since it is axisymmetric.

The problem I am encountering (second image) is that the free stream boundary is generating a non-zero velocity which increases in magnitude with increasing iterations and eventually messes up the flow field around the rotor.

I have tried increasing the domain size, increasing the mesh density at the boundary (problem is particularly bad with coarser surface mesh), I have tried a no-slip wall but nothing so far has worked.

The greatest velocity gradient is located at the corners where the free stream boundary meets the periodic boundary.

When I run this model as a steady state problem using MRF the problem was remedied by increasing the mesh density at the outer wall (free stream), but this is unfortunately not the case for when I run it as an unsteady problem using rigid body motion.

I am still an amateur in the CFD world and I am running out of ideas, and especially time..

Any help is greatly appreciated.

Regards,

D. Lehnardt

[ping]

i worry about the free stream bc use in this situation and suggest instead to use a pressure but you need to be careful due the height of your domain if gravity is enabled and then need to account for this which is well covered in a steve support article

[Rinia]

Originally Posted by D. Lehnardt

Hello everyone,

I am using Star CCM+ to model a helicopter rotor in hover flight conditions using a periodic domain of just one blade.

My boundary conditions are illustrated in the first image.

I have assigned rigid body motion to the entire domain since it is axisymmetric.

The problem I am encountering (second image) is that the free stream boundary is generating a non-zero velocity which increases in magnitude with increasing iterations and eventually messes up the flow field around the rotor.

I have tried increasing the domain size, increasing the mesh density at the boundary (problem is particularly bad with coarser surface mesh), I have tried a no-slip wall but nothing so far has worked.

The greatest velocity gradient is located at the corners where the free stream boundary meets the periodic boundary.

When I run this model as a steady state problem using MRF the problem was remedied by increasing the mesh density at the outer wall (free stream), but this is unfortunately not the case for when I run it as an unsteady problem using rigid body motion.

I am still an amateur in the CFD world and I am running out of ideas, and especially time..

Any help is greatly appreciated.

Regards,

D. Lehnardt

Dear Lehnardt,

hope you have solved the problem.

I am making a similar simulation with yours about the wind turbine on the star ccm+.

Could you give me some advice about how to use the periodic domain to perform the one blade simulation? Just set the interface to periodic?


Regrads

[SESP]

Hi David,

I would not build the case in that way.

Your approach seems fine for a steady-state analysis. It could be easier if you add stationary domain, but from your description it sounds like you got the steady case to a stable and converged state with refinement.

For the unsteady analysis with rigid body motion (or overset, which is also worth a try), I would model it as a 360 deg case. I think I never tried it, but to me the combination of rigid body motion with periodic boundary conditions does not seem like a good idea. Also here, think of adding a stationary domain around it. Personally, i would model a bigger fluid volume around the blade and then use overset on the 360 rotor. That will work without any problems, it just takes a bit of time get familiar with the setup and the relevant parameters.

Good luck,
Sebastian

[Rinia]

Originally Posted by SESP

Hi David,

I would not build the case in that way.

Your approach seems fine for a steady-state analysis. It could be easier if you add stationary domain, but from your description it sounds like you got the steady case to a stable and converged state with refinement.

For the unsteady analysis with rigid body motion (or overset, which is also worth a try), I would model it as a 360 deg case. I think I never tried it, but to me the combination of rigid body motion with periodic boundary conditions does not seem like a good idea. Also here, think of adding a stationary domain around it. Personally, i would model a bigger fluid volume around the blade and then use overset on the 360 rotor. That will work without any problems, it just takes a bit of time get familiar with the setup and the relevant parameters.

Good luck,
Sebastian

Hi, Sebastian, thanks for your reply in CFDonlien.

The UserGuide of STAR CCM+2020 provides a illustration of the use of periodic boundary in turbomachinery. I performed the simulation according to guide and found that the lift has a sudden change when the blade returns to the starting position. This may be associated to the discontinuity when the rotating region returns to the starting position. I haven't solved this problem.

I intend to study the tip vortex flow of a six-blade propeller, so only one blade is simulated to reduce the compuitational time. Do you have any better method to deal with this simulation?

[SESP]

Hi Rinia,

no, I have never tested a periodic interface with mesh motion where the mesh snaps back to its original position... In that case, the order of the interfaces (or more precisely, their index) is of importance, the interface moving must be declared first. But you probably know that already... Personally, i prefer full 360 degree when working with mesh motion for rotating components. Unfortunately, its more expensive...

Sebastian