[Kishen]

Hi,

I am trying to model a Rotor in CFD (star-ccm+). My current model is massively under predicting lift. I am not entirely sure if this is because of the initial conditions or because the mesh is not fine enough.

I have set up a moving reference frame for the rotor to which I have specified a rotational velocity. I am using the k-epsilon turbulence model for which I have calculated turbulence dissipation and kinetic energy. The Reynolds number of my rotor is based upon the rotor diameter and tip velocity and was found to be 65 million. I am also modelling it as steady state.

I am expecting around 6000N of lift force but at the moment I am only achieving 1100N. I am aware that it is difficult to accurately model a Rotor but I am sure a lot of people have achieved it. I wanted some advice on what I could do to better simulate this.

Thanks

Kishen

[Totalsim]

Have you looked at the post pro? Have the blades stalled?

[Kishen]

Yeah I've checked it and the blades haven't stalled. Also in reality my rotor isn't powered It is acting like a wind turbine. by applying a rotation to it I'm adding energy into the system rather than taking energy out. maybe I could apply a 1DOF to it or something. I have a feeling it's my model set up and initial conditions that are the problem.

I've created a much finer mesh and the load increased to 2000N. but I'm afraid that is the limit of my resources. The uni clusters won't be able to run anything better.

Ps: I've heard of TotalSim. you guys do some cool stuff

[Totalsim]

Thanks Kishen.

The relative velocity of the blade and fluid can be super critical to predicting the right forces. Have you done a speed sweep? You may only have to be 1% your blade or fluid speed to make up your deficit.

[Kishen]

Actually now that you mention it, I have been thinking about the relative velocities and this may be where I am going wrong.

What do you mean by a speed sweep? I have assigned a rotation of 250rpm to the rotor. The free stream velocity is 30m/s. Btw I am using Star-ccm+.

Best Wishes

Kishen

[Totalsim]

You can sweep either speed and it should have a similar effect.

I would tend to sweep the RPM. Maybe +/- 10%.
You then can get a power curve. It may be that only a small different in RPM will get you to your 6000N. When compared blades/turbines/propellors this is often the best way to compare different options.

Good luck.

[Alfonso]

Hi Kishen, some questions

What are the reference values of lift you are using? Theory, Wind Tunnel?

I manage years ago to obtain decent results (for lift) in 3 dimensional rotors for various angle of attack. I obtained these results in StarCCM+ with both Dynamic Meshing and MRF methods. Some comments about my experience:

- First of all. I made sure I could obtain an almost perfect cl/cd curve for the 2D airfoil.
- I made sure that my reference values was reliable and extrapolable to my case.
- Those things may sound kind of obvious but I last longer in solve those than in obtain a good 3D simulation.

Good luck