Reynolds Number

The formula and its physical meaning

The Reynolds number is the ratio of two characteristic forces acting on a moving fluid:

Re = ρvD / μ = vD / ν

  ρ — fluid density (kg/m³)
  v — characteristic velocity (m/s)
  D — characteristic length (m)
  μ — dynamic viscosity (Pa·s)
  ν — kinematic viscosity (m²/s) = μ/ρ

Inertial force per unit volume scales as ρv²/D — the rate at which a fluid parcel's momentum changes. Viscous force per unit volume scales as μv/D² — the resistance to shear set by molecular friction. The ratio of those two is exactly ρvD/μ.

When Re is small, viscosity dominates. Disturbances damp out faster than inertia can grow them. The flow is laminar. When Re is large, inertia dominates. Tiny perturbations — surface roughness, inlet swirl, vibration — get amplified rather than dissipated. The flow is turbulent.

Pipe-flow regimes at a glance

Re < 2,300    LAMINAR        ──────────────  parallel streamlines, parabolic profile
                              ──────────────
                              ──────────────

Re ≈ 2,300–4,000  TRANSITIONAL  ╲╱╲╱╲╱╲╱╲╱╲╱  intermittent puffs of turbulence
                                ╱╲╱╲╱╲╱╲╱╲╱╲

Re > 4,000     TURBULENT      ░▓▒░▓▒░▓▒░▓░  three-dimensional eddies, flat profile
                              ▓░▒▓░▒▓░▒▓░▒
                              ▒▓░▒▓░▒▓░▒▓░

Worked example: water in a 50 mm pipe

Water at room temperature has ρ = 1000 kg/m³ and μ = 10⁻³ Pa·s. A domestic plumbing line of internal diameter D = 50 mm carries water at v = 2 m/s. What's the Reynolds number, and is the flow turbulent?

Re = (1000 · 2 · 0.05) / 10⁻³
   = 100 / 10⁻³
   = 100,000

That's 40× the critical value, so yes — turbulent. The pressure-drop correlation for laminar flow (the Hagen–Poiseuille equation) does not apply; you need the Moody chart with a friction factor of about 0.03 for smooth pipes.

Drop the velocity to 0.04 m/s in the same pipe and Re = 2000. Now flow is laminar, ΔP scales linearly with v, and Hagen–Poiseuille gives an analytical answer.

Typical Reynolds numbers across fluids and scales

Scenario	Fluid	v	D	Re	Regime
Capillary blood flow	Blood	0.5 mm/s	8 μm	~10⁻³	Stokes (creeping)
Bacterium swimming	Water	30 μm/s	2 μm	~6×10⁻⁵	Stokes
Honey from a jar	Honey	10 mm/s	20 mm	~0.02	Laminar
Domestic water pipe	Water	2 m/s	50 mm	10⁵	Turbulent
Crude oil pipeline	Oil (μ≈0.05)	1 m/s	0.6 m	~10⁴	Turbulent
HVAC duct (room air)	Air	5 m/s	0.3 m	~10⁵	Turbulent
Cessna wing chord	Air	50 m/s	1.5 m	5×10⁶	Turbulent BL
747 wing chord (cruise)	Air (cold, thin)	250 m/s	8 m	5×10⁷	Turbulent BL

Notice the 12 orders of magnitude separating microbes from airliners. The same Navier–Stokes equations describe all of them — but the relative importance of inertia and viscosity is unrecognisably different at the two ends, which is why the engineering of microfluidics has nothing in common with the engineering of jet wings.

Re for water, air, and oil at common pipe sizes

Pipe D	Velocity	Water Re (ν=10⁻⁶)	Air Re (ν=1.5×10⁻⁵)	Light oil Re (ν=10⁻⁵)	Heavy oil Re (ν=10⁻⁴)
10 mm	1 m/s	10,000	667	1,000	100
25 mm	1 m/s	25,000	1,667	2,500	250
50 mm	2 m/s	100,000	6,667	10,000	1,000
100 mm	2 m/s	200,000	13,333	20,000	2,000
300 mm	3 m/s	900,000	60,000	90,000	9,000
1 m	5 m/s	5,000,000	333,000	500,000	50,000

Light oils run turbulent at industrial sizes; heavy crudes (especially cold) sit in the laminar or transitional zone. That's why heated pipelines and drag-reducing additives matter for crude transport — the regime change to turbulent dominates the pumping bill.

Reynolds similarity and wind-tunnel testing

Two flows that share the same Reynolds number — even if size, speed, and fluid differ — are dynamically similar. Streamlines look the same; pressure coefficients match; drag and lift coefficients agree. This is why a wind-tunnel model can predict full-scale aircraft behaviour, but only if Re matches.

For a 1:10 scale model, matching Re means raising the wind-tunnel velocity by 10×, or operating in pressurised air (10× density) at the same speed, or using cryogenic nitrogen (twice as dense, half the viscosity, ~5× Re-boost). NASA's National Transonic Facility uses cryogenic nitrogen at 80 K and 9 bar specifically to match flight Reynolds numbers on small models — without it, scale-model tests systematically under-predict drag because the boundary layer is too laminar.

Real-world applications

• Pipeline design. Friction factor tables (Moody chart) are organised by Re — pumping power scales linearly with friction factor at a given flow rate. A 10% mis-estimate of Re from incorrect viscosity costs 10% on the energy bill for the pipeline's life.
• Heat-exchanger sizing. Heat-transfer correlations like Dittus–Boelter (Nu = 0.023·Re⁰·⁸·Pr⁰·⁴) require turbulent flow. Designers deliberately keep Re > 10⁴ so the correlation applies and U-values stay predictable.
• Drug-delivery microfluidics. Lab-on-a-chip channels run at Re ≈ 0.01–10. Mixing must be engineered actively (chaotic advection, herringbone grooves) because turbulence is unavailable.
• Sports aerodynamics. A cricket ball has Re ≈ 10⁵ at 30 m/s; the seam triggers boundary-layer transition asymmetrically, producing reverse swing. Same physics on golf-ball dimples — they trip the boundary layer turbulent so it stays attached longer, cutting wake drag in half.
• Crude-oil pipelines. The Trans-Alaska line runs near Re ≈ 4×10⁴ at design flow. Heating the oil from 60°C to 65°C drops viscosity ~15%, raises Re, and reduces friction-factor by enough to save tens of MW of pumping.
• Sailing and rowing. A racing shell hull at Re ≈ 10⁷ has a turbulent boundary layer over almost its entire length; surface finish below 1 μm Ra is needed before further smoothing pays off.

Variants and related dimensionless numbers

• Pipe-flow Reynolds number. D = inside diameter; the canonical case with the 2300 critical value.
• Open-channel Reynolds. Uses hydraulic radius R_h = A/P and a different critical value (~500 with R_h, since R_h = D/4 for pipes).
• Plate Reynolds. Re_x = vx/ν measured along the plate; transition occurs at Re_x ≈ 5×10⁵ for smooth plates.
• Cylinder/sphere Reynolds. D = body diameter; drag coefficient drops sharply at Re ≈ 2×10⁵ (the "drag crisis") as the boundary layer trips turbulent.
• Hydraulic-diameter Reynolds. D_h = 4A/P for non-circular ducts — annular gaps, rectangular HVAC, triangular cores — extends pipe correlations approximately.
• Particle Reynolds. For settling particles, slip velocity × particle diameter; Stokes drag applies only at Re_p < 1.

The Reynolds number's siblings are all ratios of forces or transport rates: Mach (inertia/compressibility), Froude (inertia/gravity), Weber (inertia/surface tension), Prandtl (momentum/heat diffusivity), Péclet (advection/diffusion). Memorising the ratios is more useful than memorising the formulas.

Common failure modes and pitfalls

• Assuming laminar in a pipe that is actually turbulent. Hagen–Poiseuille gives the analytical pressure drop for laminar pipe flow; applied to turbulent flow it under-predicts by an order of magnitude. Many a first-year project pump comes out drastically under-spec because Re wasn't checked.
• Using the wrong characteristic length. Re_pipe (with D) and Re_plate (with x) are not interchangeable; their critical values differ by 200×. Always state which length you used.
• Ignoring temperature dependence of viscosity. Water viscosity halves between 20°C and 50°C; oil viscosity can drop 10× over the same range. A "design Re = 5000" calculated at room temperature can become Re = 50,000 in service.
• Assuming Re alone is enough. Match Re for inertial-viscous similarity; you also need to match Mach for compressibility, Froude for free-surface gravity waves, Weber for surface-tension-dominated droplets. Wind tunnels match Re; ship towing-tanks match Froude; both are wrong about the other unless you scale density and gravity creatively.
• Boundary-layer trip neglected. Sub-critical Reynolds tests on a smooth model hide the post-transition drag rise that the full-scale article will see. Wind-tunnel models often have a deliberate trip strip near the leading edge to force turbulence early.
• Roughness ignored at high Re. Above Re ~10⁶ in pipes, friction factor stops depending on Re and depends only on relative roughness ε/D. A polished pipe and a galvanised one with the same Re give very different ΔP.