Learning Forum

[Jonas4]

Hey everybody,

I simulate a stationary flow test rig with ansys fluent. The test rig measures the swirl by a paddle wheel ( tangential over axial velocity) and the flow coefficient with a flow meter (measured mass flow over theoretical mass flow).

My question is now how i can calculate a swirl number in a plane for my simulation and the flow coefficient at my outlet?

 

Thanks in advance for your help!

[Amine Ben Hadj Ali]

Have a look into /forum/forums/topic/swirl-number-calculation-in-fluent/

[Jonas4]

Thanks, but somehow i can't find the definition of the swirl number in Fluent under Turbulence modeling...

I thought about calculating it afterwards in the results, how is this possible there?

[Amine Ben Hadj Ali]

 Chapter 9.3.2 User's Guide. You define the function accordingly or just according to the formula you have. Or just stick to the "weird" but practical formula of just having the ratio of angular velocity / axial velocity to get the swirl quantified.

[Jonas4]

I managed now to calculate the flow coefficient and could verify it with the test bench.

For the swirl i need some further help please. I would like to calculate the swirl in a plane. The swirl at the test bench is defined as tangential velocity over axial velocity. I now use the area weigthed average in the defined plane for axial and tangential velocity and calculate with that the swirl. The results deviate from the test bench. Is the way i calculate it wrong or is there a better one?

 

[Amine Ben Hadj Ali]

Better to use mass flow weighted averaging. 

There are a lot of methods to calculate the Swirl number, check the method used in the reference you compare to.

[Jonas4]

where do i find the option mass flow weighted averaging?

The axial velocity fits quiet good to the measurement. The tangential velocity is the problem, it's calculated by the speed of the paddle wheel of the test rig c_tang=2*n*r_paddle*pi

[Amine Ben Hadj Ali]

Is named mass weighted under surface integrals.

Perhaps you need better turbulence model which is more sensitive to streamline curvature. You can either switch on curvature correction or use high 2nd moment closure models if you want to remain RANS.

[Jonas4]

Okay thanks for that!

One more question: How can i create a plane with a hole inside, on which i can evaluate my results then. So not creating a standard plane but one with a origin and radius which is not included for the plane.

 

Best regards,

Jonas

[Rob]

You may need a custom field function that gives you radius. Then look at the iso-clip options in the surface create/manage tools.