Laminar vs Turbulent Flow — the Reynolds Number

Picture a fluid element marching down a pipe. Two influences compete for its fate. Its inertia — proportional to ρV² — is its tendency to keep moving in whatever direction it happens to be going, overshooting, swinging wide, amplifying any little swirl it picks up. Its viscosity — proportional to μV/D — is the internal friction that damps those swirls out, smoothing the motion into tidy layers. Take the ratio of the two and the velocity cancels in a pleasing way, leaving a dimensionless number that depends on density, speed, size, and viscosity.

That ratio is the Reynolds number, named for Osborne Reynolds, who in 1883 injected a thin thread of dye into water flowing through a glass pipe and watched it: a straight, unbroken streak at low flow (laminar), and a suddenly erupting, chaotic cloud at higher flow (turbulent). The switch always happened at the same value of the ratio, whatever the individual values of speed, diameter, or fluid. That is the power of a dimensionless group — it collapses a whole family of experiments onto one number.

For flow inside a smooth circular pipe, experience fixes three broad bands. Below Re ≈ 2300 the flow is almost always laminar: the dye thread stays straight, the fluid moves in smooth concentric layers, and there is no mixing across the pipe. Above Re ≈ 4000 it is almost always turbulent: chaotic eddies mix the fluid across the whole cross-section, the dye erupts, and the velocity at any one point fluctuates wildly about a mean. Between 2300 and 4000 lies the transitional band, where the flow flickers between the two and is genuinely unpredictable.

The number 2300 is not magic and it is not exact. It depends on how disturbance-free the inlet is, how rough the pipe is, and on tiny vibrations. In an exceptionally quiet laboratory setup laminar flow has been held to Re of tens of thousands; in a noisy industrial pipe the transition can begin below 2000. For engineering work, treat 2300 as a practical dividing line and the 2300-4000 band as a region where you should not trust a single friction-factor number.

The profile shape in the diagram is the visible symptom of a deeper difference. In laminar flow, momentum crosses the pipe only by viscous diffusion — slow, molecular, layer-by-layer. The result is the sharp parabolic profile, with fluid at the centreline travelling twice as fast as the mean and fluid at the wall stuck fast by the no-slip condition. There is no mixing of the bulk fluid at all.

In turbulent flow, eddies — macroscopic swirling globs of fluid — fling themselves across the pipe, carrying fast core fluid to the wall and slow wall fluid to the core. This turbulent mixing is orders of magnitude faster than viscous diffusion, so it flattens the profile: the centreline speed drops closer to the mean (the ratio V/u_max rises from about 0.5 in laminar flow to about 0.8 in turbulent flow), while the velocity near the wall rises. That steeper near-wall gradient is why turbulent flow has a much larger wall shear stress and therefore a larger pressure drop for the same flow rate — the central reason pumping turbulent flow costs more energy.

Fluid entering a pipe does not arrive with its fully developed profile. At the inlet the velocity is roughly uniform (a flat front); then viscosity, working inward from the no-slip wall, gradually builds up the profile over an entrance length L_e. Only beyond that length is the flow fully developed — the profile stops changing with distance, and the clean formulas of the next lessons apply.

For laminar flow the entrance length is surprisingly long, scaling linearly with Reynolds number. For turbulent flow the eddies mix the profile into shape much faster, so the turbulent entrance length is shorter and nearly independent of Reynolds number. In most long-pipe problems the entrance is a small fraction of the total length and is neglected; in short pipes or capillaries it matters.