Is the tunnel just a pretty picture, or does it get numbers right? These are the standard test cases for a circular cylinder: drag coefficient in the steady regime, and drag plus vortex-shedding frequency (Strouhal number) once the wake becomes unsteady. Each case runs here in your browser, on a cylinder 32 cells across in a tunnel 16 diameters tall, and is compared with values from the literature.
| Case | Grid | CD | CD reference | Error | St | St Williamson | Error | Status |
|---|
What is being compared. Reference drag values come from the experimental and numerical literature for an unbounded two-dimensional cylinder (Tritton 1959; Henderson 1995; Williamson 1996), given as a central value and the spread between sources. The Strouhal reference is Williamson's 1988 relation St = 0.1816 − 3.3265/Re + 1.6·10⁻⁴·Re, which fits careful experiments in the laminar shedding regime to within about a percent.
Where differences come from. The tunnel is 16 diameters tall with slip walls. Confining a cylinder like that is known to raise its drag by very roughly 5–10% relative to an unbounded flow at these Reynolds numbers, and the bounce-back boundary sits half a cell outside the nominal circle, which adds about 3% more. So expect drag to come out several percent high; the Strouhal number is far less sensitive to confinement and is the sharper test. The lattice Mach number is 0.14, so the fluid is slightly compressible. The cylinder is a staircase of cells, which matters most at 16 cells across; the grid check shows how much. Each unsteady case is given 140 diameters of travel to settle after a short spin of the cylinder that breaks the symmetry, then 70 diameters of measurement, so the shedding frequency is resolved from a few dozen cycles.
What this does not test. Nothing here checks the sub-grid turbulence model, the wing at high Reynolds number, or the moving-body force term beyond the symmetry-breaking kick. Those regimes are shown in the tunnel as qualitatively right, not validated.