[Marco D'Ario]

Hi, I have been using SU2_CFD for a couple of months and I am interested in the CD of small irregular particles (0.05 < Re < 5000, fluid velocity << 100 m/s).

I am interested in the transition between laminar and turbulent flow so I have used the INC_RANS solver for an incompressible fluid.

With my set-up, I can calculate the CD of a sphere with 0.05 < Re < 1000 with an error smaller than 3% and smaller than 10% for Re > 1000.

However, when simulating irregular objects the CD is always smaller than the sphere for Re < 200. This made me realize that the CD I obtain (I get it from the Hustory output) must be calculated using some area for an equivalent sphere, either the projected area of an equivalent sphere or the surface area of an equivalent sphere. When I account for that the CD obtained is closer to the one expected for my particles.

Does anyone know the exact what that SU2 calculates the CD? What do normally users do with the CD obtained from SU2 (assuming they use the CD from the History output file)? Or is there a way to get the "actual CD" for the simulation? Alternatively, are any of my assumptions wrong?

Thanks so much for your help.

[LuckyTran]

From a practical standpoint people do exactly what you have done. They experiment with different areas or just pick a convenient one. At the end of the day, none of it matters as long you consistently reuse the same reference areas and Reynolds numbers.


From a theoretical standpoint, you are supposed to simulate the case the case of a sphere at the exact same condition and take the ratio of drag forces for your arbitrarily complex object and the spherical. However, nobody has the time to simulate a sphere when they want to spend their time simulating the flight of dinosaurs on the planet Mars. Hence, no one actually calculates a proper Cd, ever.

[Marco D'Ario]

Thanks, LuckyTran for your reply.

I would like to compare how ornamentation affects the aerodynamics of pollen grains, so I would like to have a good approximation of the CD. I must say, I am surprised that the obtained CD is basically "just a number" for me to play around with.

As I said, the simulations of my sphere are close to known values. Would you recommend tuning my CD on known objects? If so, can anyone suggest known standards for calibrations? I know oblate spheroids and octahedra are often used.

Thanks again!

IMPORTANTLY: I would like to clarify from my query that the history output always gives me a CD lower than the CD I obtain from a sphere, for Re<100. Only when in normalize by the area do I get more reasonable CDs. And yet, I have symmetric concave shapes in which the CD is lower than that of a sphere (symmetric, NOT like an ellipsoid normal to the wind, for instance.). I of course compare shapes ar similar Re.

In my understanding, irregular particles are expected to have a CD higher than the one of a sphere, so that is why those results are puzzling to me.