CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Quantifying swirl (

Nak December 11, 2006 17:00

Quantifying swirl
I suspect some amount of swirl in the flow on the outlet face of my pipe-geometry that I solved in FLUENT. I can see this from a plot of velocity vectors on that plane. Can someone suggest me a method to quantify this swirl and also create contour plots?

Jason December 11, 2006 17:36

Re: Quantifying swirl
Set the axis of your fluid so it passes through the center of this face (if it doesn't already... you can do this in Define->Boundary Conditions and select the fluid volume) and then you can look at tangential velocity. That'll work for contour plots. For quantifying, what are you trying to get across? You could do an average of the tangential velocity in the Report->Surface Integrals.

Hope this helps, and good luck, Jason

Nak December 11, 2006 19:18

Re: Quantifying swirl
Thanks for the response Jason. Your suggestion of setting the axis of fluid seems to work. For quantifying, I figured out to express that swirl in terms of an "angle": arctan(v-tangential/v-axial)

I guess I have two methods to do this: Method1:Should I write a custom-field function for this eqn and then get an area-wtd average, or Method2: just compute average v-tangential and average v-axial, divide them and then take inverse tangent. Hope I am making sense... Which one would be correct? Or do you have something else?

Thanks in advance...


hugo December 12, 2006 09:58

Re: Quantifying swirl

suppose that:

U -> axial velocity. V -> circunferencial velocity. r -> radial coordinate. tita -> angular coordinate R -> the pipe radio.

now suppose that:

A -> surface integral between 0 and R in the radial coordinate and between 0 and 2Pi in the angular coordinate of the quantity U*V*r in the interest cross section of the duct.

B -> surface integral between 0 and R in the radial coordinate and between o and 2Pi in the angular coordinate of the quantity (U^2)*R in the interest cross section of the duct.

The deficition of Swirl of Gupta (1984) is:

Sn = A/B

You can define the quantity U*V*r and the quantity (U^2)*R in: Define>Custom Field Functions

The in Report>Surface Integral you can calculate A and B

sorry for my English


Jason December 12, 2006 10:03

Re: Quantifying swirl
You can try it both ways (they seem relatively simple to try), but I'd use a custom field function. Using a CFF would also allow you to plot it as a contour.

Good luck, Jason

Nak December 12, 2006 14:25

Re: Quantifying swirl
OK, I calculated the Gupta-definition of Swirl number using custom field functions (one each for tangential and axial fluxes and then divide the numbers).

Question: I get a negative value, -0.02, is it OK to get a negative swirl number? Does make sense physically?

Thank you all for the inputs...


hugo December 12, 2006 15:37

Re: Quantifying swirl
Nak. if you check The definition that I gave you, B is always positive but A is not necessary positive and the sign depend on the direction of the circunferencial velocity.

It's Ok to get a negative swirl number.

good luck

All times are GMT -4. The time now is 10:22.