[Panee]

Dear all,

I am new to fluent and I am trying to simulate a two bladed Savonius rotor.

I am using a sliding mesh to simulate the turbine rotation and the angular velocity is set to 20/3 radians per second. I am using a realizable k- epsilon turbulence model with standard wall function. A time step of 0.0013 seconds corresponding to 0.5 degree rotation is being used. Convergent criteria for all variables is set to 1e-5. I am generating the mesh in ANSYS itself.

The torque coefficient variation is expected to be sinusoidal in nature. However the torque coefficient variation is sinusoidal with lots of noise added to it. I have posted the image of the output as well as the mesh for your consideration. Kindly let me know what the problem with the simulation is. I would be happy to furnish more information if required.

Note - The simulation used first order upwinding for the calculation of all variables. Second order upwinding gave poorer results.