[Rockytime]

Hi All,

I failed to get my force value at my turbine blade surface.
It show force value = 0N. FYI, I need to use the force value to calculate the torque of the turbine. Can anyone advise me on how to get the force value?

For my CFX-pre, I set

1) Inlet = (30*sin((2*pi*t)/WvePeriod)) [m s^-1]
2) Rotation Velocity of turbine = 100rpm
3) Frce = force_x()@REGION:BladeSurface

Picture below show the blade surface.


Picture below show the force = 0N. Variable 1 = Force


If you guys need more information, attached below is the CEL coding. Your help will be much appreciate. Thank you.

[ghorrocks]

Lots of comments:

Why have you turned buoyancy on? It does not appear to contribute.

Have you modelled the rotor as a solid body? Why? Are you doing FSI or rigid body dynamics? If not then do not model the rotor as a solid.

Why are you modelling this as transient? Should this be a steady state run?

Force calculations on interfaces can be tricky. You need to put it on the correct side of the interface. But as I suspect you don't need an interface when you fix this the problem will magically resolve itself.

[Rockytime]

Hi Glen,

Thanks for your reply.

1) The reason I set buoyancy is because I need the gravity value on the Y-direction . My turbine duct is set up in a vertical direction (Y-direction). Hence, I think I need to set the gravity direction which can be found in the buoyancy settings.

2) I need to study both fluid structure interaction (FSI) and rigid body dynamics.
First, I need to study the separation flow that form behind the turbine when the fluid flow through.
Second, I need to study how much torque can be produced in by the air that flow through the turbine. Hence, I need to know how much force exert by the air flow on the blade surface in order to calculate the torque.

3) I got ask some of my friends, they say is better to use transient but after read through your explanation in previous thread. I think is ok for me to use steady state as long as the result run to final convergence.

I have remove the interface, buoyancy and try to run the result in steady state again. I not sure why it show 0 value on the FluidBody. Appreciate if you can guide and explain to me on this. Attached file below is the CEL coding for your further understanding about my problem.

[ghorrocks]

The zero in FluidBody usually means that there is nothing to move the fluid in that domain, so it just stays stationary. It usually means you forgot to apply boundary conditions to that domain or you did not link it to the rest of the flow with an interface properly.

[Rockytime]

Hi Glen,

I did apply boundary condition on my fluid body, which is the inlet, opening and wall boundary condition.

But I'm confused the setup for interfaces and domain for turbine.

1)If I need to investigate the separation flow that occurs behind the turbine and calculate the force exert by the air flow on the blade surface. Can I consider this 2 problem as FSI and rigid body dynamics? If yes, then is it ok if I continue to model my rotor as solid body?

2) Regarding interface setup, I'm choosing (general connection between the Fluid Body and Rotor), (frozon rotor with 0 degree of rotational offset). Am I doing it in the right way? If no, can you please advise me. Appreciate for your help.

[ghorrocks]

FSI is for where the fluid acts on the solid and deforms it (or effects the solid is some other way). If the solid is not deformed by the fluid then the simulation is not FSI.

Rigid Body is for where the forces on the body cause the body to move in a motion which is not known beforehand.

If the motion is known beforehand - which is usually the case for rotor modelling - then it is either moving mesh or rotating frame of reference.

So most rotor models are just rotating frame of reference. No FSI, no rigid body dynamics.

Whether the flow separates or not is irrelevant to the choice of physical model. All these models can predict a separated flow and the forces which it generates.

Can you please show an image which shows your stationary and rotating domains and where you put the interface.

[Rockytime]

Hi Glen,

Apologise for taking a long time to reply. I'm trying out myself after read your comments. There is another problem after I switch my rotor models to fluid domain. The result show symmetry for front and back of the rotor which I think is wrong.





Besides that, the solver manager keep show "A wall has been placed at portion(s) of an OUTLET boundary condition (at 31.6% of the faces, 31.2% of the area) " and recommend me to change to opening and I tried before and the result still showing symmetry result.






And I have another question which is varying my inlet velocity. How can I vary my inlet velocity in steady state?
Is it corret if I set Inlet2 = 0.5 [m s^-2]*t ,so the velocity will increase when the timestep increase.



Below is the picture of stationary Domain, rotating domain and interface.


Stationary




Rotating




Interface


Attached below is the CEL Code for further information. Appreciate your help, and sorry in advance because I might be asking some basic questions.