How this instrument works
Re weighs a fluid's inertia against its own stickiness. Momentum carried by moving mass sits on top of that fraction; viscous shear, which smears velocity differences flat, sits underneath. A large ratio means inertia wins and any small wobble grows into eddies. A small ratio means viscosity damps disturbances before they can organise, so layers slide past one another in orderly sheets.
George Stokes wrote this grouping down in 1851, but Reynolds gave it teeth at Manchester in 1883 by injecting a dye filament into water and watching it break, at one definite speed, into sinuous chaos. Sommerfeld attached Reynolds' name to it in 1908. Pipe thresholds quoted everywhere descend directly from that apparatus: below roughly 2300 flow stays laminar, above about 4000 it is dependably turbulent, and territory between behaves like whichever it fancies on a given morning.
None of this is universal, because characteristic length D gets chosen by convention rather than derived from theory. Round pipe flow uses internal bore, spheres use their diameter, flat plates use distance from their leading edge, and rectangular ducts want hydraulic diameter — four times area over wetted perimeter. Shift that convention and every critical threshold shifts too, which is why a Reynolds number quoted without its length scale is only half measured.
- Enter Fluid density in kg/m3 — near 998 for water at room temperature, about 1.2 for air at sea level.
- Give Flow speed as a bulk average: volumetric throughput divided by cross-section, not that faster centreline peak.
- Set Pipe diameter to internal bore. For any non-circular duct, substitute hydraulic diameter instead.
- Enter Dynamic viscosity (Pa·s). Divide centipoise by 1000; water at 20 °C lands on 0.001 Pa·s.
- Read Reynolds number and place it: under 2300 laminar, over 4000 turbulent, anything between unsettled.
Worked example — water at 2 m/s down a 50 mm main
Cold water, density 1000 kg/m³ and dynamic viscosity 0.001 Pa·s, moving at 2 m/s along a 50 mm bore. Substituting: Re = 1000 × 2 × 0.05 ⁄ 0.001 = 100 ⁄ 0.001 = 100000. Nothing rounds anywhere; arithmetic lands exactly on one hundred thousand.
At 100000 you sit far past 4000, so expect a fully turbulent core, a thin viscous sublayer clinging to metal, and friction scaling closer to speed squared than to speed itself. Two practical consequences: pressure loss climbs steeply should a bigger pump tempt anyone toward 3 m/s, while wall heat transfer improves markedly for that very same reason — mixing.
Questions
Should I enter dynamic or kinematic viscosity?
Dynamic, in pascal-seconds. Kinematic viscosity ν already has density folded in, so pairing it with a density field double-counts and inflates Re by roughly a thousand for water. Convert first: 1 centipoise = 0.001 Pa·s, 1 poise = 0.1 Pa·s. If your datasheet quotes centistokes instead, multiply by density in kg/m³ and divide by a million to reach Pa·s.
What counts as turbulent?
For round pipes, laminar below about 2300 and turbulent above roughly 4000, with a grudging transitional band between where very smooth inlets hold laminar flow far higher than textbooks suggest. Those figures belong to pipes alone. Flow over flat plates trips near 500000 measured from the leading edge, while spheres keep their laminar boundary layer until about 300000, where drag suddenly collapses — dimpled golf balls exploit exactly that crisis.
Why does a dimensionless number matter so much?
Because two flows sharing one Reynolds number behave identically in shape, whatever their physical size. That principle, dynamic similarity, is what lets a scale model in a wind tunnel stand in for a full aircraft. Matching Re on a small model demands high speed, dense gas or low temperature, which is precisely why facilities such as NASA's National Transonic Facility pressurise and chill their nitrogen.
How do I handle a square duct or an open channel?
Swap in hydraulic diameter: four times flow area divided by wetted perimeter. Square ducts of side s return exactly s. Wide open channels of depth h return roughly 4h, since only bed and banks count as wetted — free surface does not. Feed that figure into Pipe diameter and read Re as normal, though critical thresholds for open channels differ from pipe values.
Does Reynolds number give me pressure drop directly?
No. It selects a friction factor, which then produces pressure drop through the Darcy-Weisbach equation. Below 2300 that factor is simply 64/Re. Above transition it depends on Re and relative wall roughness together, read off a Moody chart or solved from Colebrook. Skipping roughness in turbulent pipe flow is a routine source of undersized pumps.
What does life look like at very low Reynolds number?
Bizarre. A swimming bacterium runs near 0.00001, where viscosity so dominates that coasting is impossible — stop pushing and it halts within an atom's width. Edward Purcell's 1977 lecture framed the consequence: any reciprocal stroke, a scallop opening and shutting, produces zero net travel. Hence corkscrew flagella. Blood in capillaries and oil creeping through rock sit in similar territory, well under 1.