Sphere Drag Against the Stokes and Oseen Solutions — Troubleshooting Guide

Sphere Drag as a Validation Benchmark

Drag on a sphere in uniform flow is the analytical-solution benchmark used most often to validate CFD accuracy. The reference theory comes in three tiers.

Here \( Re = \rho U d / \mu \) (diameter-based) and \( C_D = F_D / (\frac{1}{2}\rho U^2 \cdot \frac{\pi d^2}{4}) \) (projected-area-based). This page works through the causes of "my CFD sphere will not match theory", in order of how often they turn out to be the culprit.

Symptom 0: Mismatched Definitions (Three Minutes, Before Anything Else)

• Length scale in Re — diameter or radius? Building Re on the radius puts you systematically off the theory by a factor of 2. Are the literature, the tool and your own post-processing all using the same one?
• Reference area — \( C_D \) is referred to the projected area \( \pi d^2/4 \). Dividing by the surface area \( \pi d^2 \) is off by a factor of 4
• Drag components — is it pressure drag plus friction drag? At low Re friction accounts for two thirds of the total (the Stokes solution splits 1/3 pressure, 2/3 friction), so picking up only one of them misses badly
• Reference velocity — the free-stream velocity, or the blockage-corrected velocity (when comparing against a wind tunnel)?

When the discrepancy is a clean ratio — double, half, 30% — the cause is neither physics nor numerics. It is almost always a definition mismatch.

Symptom 1: Drag Above Theory at Low Re — Blockage From the Domain Size

Because a Stokes-flow disturbance decays so slowly (\( u' \sim 1/r \)), the lateral boundaries of a cylindrical or prismatic domain act as invisible walls and inflate the drag. In the classical wall corrections (the Bohlin/Haberman correction for a sphere in a tube), the correction enters as \( 1/(1 - 2.104\lambda + \dots) \) in terms of the sphere-to-tube diameter ratio \( \lambda = d/D \), giving about a 4% increase even at \( \lambda = 0.02 \) (a domain 50 times the sphere diameter). Three options are open in practice.

1. Make the domain large enough — for 1% target accuracy, a radial extent of at least 100 sphere diameters is the guide. For a low-Re benchmark, a domain that feels too big is the correct one
2. Correct the theory side with a wall-correction formula — compare against the finite-domain theoretical value (Stokes drag with the wall correction). A smart way to sidestep the cost of enlarging the domain
3. Run a domain-size study — compute the drag at 25D/50D/100D and extrapolate to the infinite-domain value using the \( 1/r \) behaviour. It doubles as the diagnostic for whether blockage is really the cause

Symptom 2: "It Does Not Match the Stokes Solution" at Re = 1 to 10

It is not supposed to match. The error of the Stokes solution already exceeds 10% at Re=1 and grows to a factor of several by Re=10 — this is picking the wrong reference to compare against, not a defect in the CFD. Above Re=0.1 use the Oseen correction, and above Re=1 the standard drag curve (the empirical correlation). Turned around, the two ways to use the sphere as a CFD validation case are ① compare against the Stokes solution at Re≪1 (a precise check on the diffusion terms and the boundary treatment) and ② compare against the standard curve at Re = 10 to 100 (a check on the convection terms and the wake). Run as a pair, they exercise different sides of the code. The correlations themselves scatter by a few percent, so use ±5-10% as the pass/fail band for Re>1.

Symptom 3: Convergence Fails or the Answer Wanders as Re Rises — the Flow Regime Has Changed

"My steady analysis will not converge at Re=300" is not a sick solver but the physical answer that no steady solution exists (the general rule from convergence diagnosis). Likewise, when an axisymmetric analysis at Re=250 "converges but disagrees with experiment", the enforced symmetry has manufactured a nonphysical solution. Check the map from Re range to analysis type before you do anything else.

Symptom 4: Refining the Mesh Does Not Settle the Answer, or the Value Drifts

• Resolution near the sphere surface — the drag is an integral of surface pressure and shear. That demands enough layers through the boundary layer (at low Re, a viscous layer on the order of the sphere diameter) and a smooth discretization of the sphere itself (a coarse polyhedral approximation feeds surface-area and curvature error straight into the result)
• Wake resolution — above Re=20, under-resolving the separated region distorts the pressure drag. Do not forget to refine downstream
• Convergence criterion — low Re with a large domain converges slowly. Keep iterating until the drag history is completely flat (never stop on residuals alone)
• Boundary conditions — the lateral boundaries are better treated as far-field inflow/opening than as slip walls (slip walls still leave blockage behind). Put the outlet well downstream of the sphere, and note that at low Re the upstream side matters just as much, since Stokes flow is nearly fore-aft symmetric

A Recommended Validation Protocol (One Day, in This Order)

1. Definition check — write down the definitions of Re, reference area and force components (the checklist under Symptom 0)
2. Re=0.05, domain 100D — target the Stokes solution to within 1-2%. If you miss, isolate blockage with a domain study (25/50/100D)
3. Mesh convergence — GCI on three levels (procedure). Drag is an integrated quantity and converges relatively well
4. Re=100 — standard drag curve to within ±5%. Confirm the effect of wake refinement
5. Record it — keep the domain size, mesh, convergence criteria and comparison basis (the formula) in the validation log, and adopt them as the standard settings for later projects

Related: index of analytical benchmarks, GCI troubleshooting, checking and diagnosing solver convergence.