Hi,

I'm a beginner user of CFX 5.5 and I am having the following doubt: In a duct curve (fluid = Air) there is a increase of velocity in the internal part of the curve( in the fold), opposing experimental studies. Somebody know what I am making wrong. Somebody had the same problem?

Thank you

Hi Christian,

This is exactly correct and does not oppose experimental studies. The short answer is that the flow-field is essentially an inviscid vortex at the corner, with a higher velocity at the center and a lower velocity towards the outside.

One way to think of it is to consider what happens as the fluid goes around the curve in the duct. In order for that element of fluid to turn, an inward pressure must be applied. That static pressure increases as you move to the outer part of the curve (or, inversely it must decrease as you approach from the outer part of the curve and is balanced by an increase in momentum). Since the total pressure must remain constant, the dynamic pressure adjusts to accomodate it. At the inside corner, the static pressure is the lowest and thus the dynamic is the highest. Hence the velocity is greatest at the inside corner.

At first glance this may be counter intuitive. If you watch water flowing in a stream, it will tend to go to the outside of the corner, not the inside. You may also have some other experiemental evidence where this occurs. The difference in these cases relates to the boundary layer. The total pressure of fluid in the boundary layer is lower than that of the free-stream fluid. As the thick boundary layer approaches the curve in your duct, the total pressure may not be sufficient for the fluid to accelerate at the corner and the boundary layer will separate as the higher momentum fluid is driven to the ouside of the corner.

Add a boundary layer profile to your inlet (or extend the inlet further upstream) and you should see this occur.

By the way, CFX-5.5.1 was released almost a year ago and you should not be using CFX-5.5 (not to mention CFX-5.6 is about a month away).

Regards, Robin

Robin, Thank you!

You are right, but I try change the fluid to water (and after other fluids) and the result were the same (exactly). What are we doing wrong ?

I am using the CFX 5.5.1.

Thank you again.

Best Regards

Christian Cordeiro

Hi Christian,

Although different fluid look the same, compare the actual values of pressure and you should see differences. Extending the inlet farther upstream, or adding a boundary layer profile to your inlet boundary condition will be likely to cause separation.

By the way, it is the same effect which generates lift over an airfoil.

Regards, Robin