[seand]

I am attempting to simulate flow over a cylinder at Re = 20, and compare it to some published data. The fluent tutorial for steady laminar flow over cylinder uses Forberg's research as a basis for validation, so I was hoping to do the same. I am using a similar domain shape as the tutorial, as well as domain-to-cylinder-diameter ratio. Fluid properties are slightly modified but kept constant.

Although I am looking at the same reynolds number, I am using a different diameter and velocity than the tutorial. When i attempt to plot the pressure distribution in the domain, I get a max pressure value that is significantly smaller (~0.013 Pa) than the results in the tutorial and Fornberg's paper (~0.6 Pa). I was wondering what the cause of this might be? As long as the reynolds number is the same I should be obtaining the same max magnitude pressure right? The pressure distributions along the circumference have the exact same pattern, its just orders of magnitudes smaller. Is this a fluent thing or a fundamental misunderstanding of the dimensionless parameters?

[LuckyTran]

Originally Posted by seand

As long as the reynolds number is the same I should be obtaining the same max magnitude pressure right? The pressure distributions along the circumference have the exact same pattern, its just orders of magnitudes smaller. Is this a fluent thing or a fundamental misunderstanding of the dimensionless parameters?

If the Reynolds number is the same, you get the same non-dimensional pressure. You need to non-dimensionalize the pressure and compare them.

[seand]

I was thinking this might have been dealing with scaling issues. Here's why I'm asking though: my real concern is the drag coefficient. I am plotting my Cd convergence, and it is coming out to nearly 0. (below 0.005). Where tables such as :[https://www.princeton.edu/~asmits/Bi...efficient.GIF]
show that I should be closer to 4. I assumed this was related to the smaller scale pressure distribution along the circumference of the much smaller cylinder I used. But I'm confused because I thought that was the point of using dimensionless numbers, so that tables like the one I linked to would be applicable to a wide variety of velocity/diameter configurations, as long as reynolds numbers were kept the same? If this IS the case, why does changing velocity/diameter configurations, but keeping the reynolds number the same get me different converged Cd values?

[LuckyTran]

You are on the right track. Double, triple, quadruple check the drag coefficient calculation in Fluent to make sure it is correct.

Btw, did you remember to set all the reference values?

[seand]

I modified my reference area (equal to my cylinder diameter), and I obtained a value ~2. Matches Fornberg's Cd calculation at Re 20, but looking at some Cd Vs Re tables on google, for a smooth cylinder, I'm realizing how inconsistent some of these tables are. My textbook table for Cd vs Re for a smooth cylinder shows I should be getting a value a little over 5 at Re = 20. This is something I might have to ask a fluids professor in person, i think.

Regardless, thank you for your help Luckytran