CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Cd problems (

Chetan Kadakia July 7, 2001 19:16

Cd problems
I'm running both laminar and turbulent flow past a cylinder. I monitor coefficient of drag on the cylinder surface and where my force vector is x=1, y=0. In the laminar case I get a Cd much smaller than expected and on the turbulent case I am getting a Cd much higher than expected. I am using Cd vs. Re for a cylinder published data. Am I using the force monitoring incorrectly? Please advise.

Alain July 8, 2001 08:53

Re: Cd problems
Do you have a mesh independant solution ?

I think that this case is one of the validatio cases of Fluent. So I should find more information on Fluent's CDs

Best regards


John C. Chien July 8, 2001 12:55

Re: Cd problems
(1). When you have problems , you can try simple problems such as "flow over a flat plate". (2). Everyone who is using a commercial should try to obtain the solution to the flow over a flat plate first. In this way, it will verify your skill and the code's capability, or the combination. (2). If you can't get accurate solution for the flow over a flat plate, then you should not be doing CFD at all. Because the only other simpler case is the uniform flow. (3). And you would find that even for the simple flow over the flat plate, to get accurate solution is also very challenging. (I could not even pass this step when checking out the Fluent code. As I changed the mesh, my flat plate solution also changed)

Chetan Kadakia July 8, 2001 13:46

Re: Cd problems
I have a grid independent solution for the laminar case, that I know. I have done the flat plate with confidence many months back. Any other suggestions?

Chetan Kadakia July 8, 2001 13:59

Re: Cd problems
One more thing: when I perform the unsteady laminar solution, I get an appropriate frequency (1/time period) and my seperation point and bondary layer thicknesses (momentum and displacement) all are within close approximation to values obtain using Thwaites approximation. It leads me to believe I have a correct solution for the most part. And I would think drag being viscous and not induced (the steady lift value is and should be zero) would be a primary function of accuracy within the boundary layer. My grid is also grid independent with respect to the boundary layer (as well as for fresstream conditions). Could it have to do with my use of velocity inlet and outlow conditions? My choice of constant density (fluid is a water), a use of 1st order momentum discretization (using 2nd order here would not allow for convergence)? Am I supposed to supply a reference value for reporting (I did not do that)? Let me know if you think of any thing.

John C. Chien July 8, 2001 18:45

Re: Cd problems
(1). If you are dealing with laminar flows, then the only source of inaccuracy will be the algorithm used in the code. (2). You can select higher-order algorithms available in the code, and see whether you are getting better results. (3). What is the mesh size and the Reynolds number of your case?

Bipin July 30, 2001 08:31

Re: Cd problems
Did you change the reference values like the velocity, density, area etc. in the Fluent panel

Go to Report ----> Reference Values and change the values. You will need to change area and velocity.

John C. Chien July 30, 2001 12:59

Re: Cd problems
(1). This is one of the black hole for me a few years ago. (2). I think, it is hard to fix the GUI, so that the user can input the reference values before computing the coefficients.

All times are GMT -4. The time now is 18:30.