How this instrument works
The Nusselt number compares two ways heat could cross a fluid layer of thickness L: the way it actually happens, with the fluid in motion, against the way it would happen if the fluid sat perfectly still and heat crept across by conduction alone. A stagnant film of thickness L conducts heat with an effective coefficient of k ⁄ L, straight from Fourier's law. Divide the real convective coefficient h by that baseline and the length cancels neatly into hL ⁄ k — a single dimensionless number stating how many times better convection is doing the job than conduction would.
Wilhelm Nusselt introduced the ratio in 1915, working out how boundary-layer heat transfer could be reduced to correlations useful to engineers rather than re-solved from scratch for every geometry. A Nusselt number of 1 means convection is contributing nothing beyond plain conduction, as if the fluid were frozen solid. Turbulent water in a pipe routinely reaches Nu in the hundreds; gentle natural convection off a warm wall in still air might sit in the tens. The number itself does not predict h; it is what results once h has already been measured or pulled from a correlation, such as Dittus-Boelter for pipe flow or Churchill-Chu for a vertical plate.
The characteristic length is not a free choice — whichever length defined the correlation that produced h, be it pipe diameter, plate length, or sphere diameter, is the one that belongs here, because both lengths need to describe the same physical situation for the ratio to mean anything. Swap in an unrelated length and the arithmetic still runs, but the result no longer represents the flow being studied, and that mismatch is the one mistake this instrument cannot catch on your behalf.
- Enter the convective heat transfer coefficient in the Convective heat transfer coefficient field — pull it from a correlation such as Dittus-Boelter or from measured data.
- Enter the Characteristic length: the pipe diameter, plate length, or sphere diameter that h was calculated against, in metres or centimetres.
- Enter the fluid's value in the Thermal conductivity of the fluid field, evaluated at the fluid's bulk or film temperature, not the solid surface's.
- Read the Nusselt number result — values near 1 mean conduction is doing almost all the work; higher values mean convection dominates.
Worked example — a 0.5 m plate in slow-moving water
Take a submerged plate with a characteristic length of 0.5 m, a convective heat transfer coefficient of 25 W/(m²·K) — plausible for gentle natural convection in water — and a fluid thermal conductivity of 0.6 W/(m·K), close to water's actual value near room temperature. The formula gives Nu = 25 × 0.5 ⁄ 0.6, which is 12.5 ⁄ 0.6, or exactly 125 ⁄ 6 = 20.8333333333, a repeating decimal that never resolves cleanly.
That figure, just under 21, says the moving water is transporting roughly twenty-one times more heat across the plate's boundary layer than a stagnant 0.5 m film of the same water would by conduction alone. Double the coefficient to 50 W/(m²·K) with everything else fixed and Nu doubles to about 41.67, because the formula is linear in h; double the length to 1.0 m instead and Nu reaches that same 41.67, because it is equally linear in L.
Questions
What does a Nusselt number of about 21 actually mean?
It means the moving fluid is transferring roughly 21 times more heat across the boundary layer than a motionless film of the same thickness would by conduction alone. Nusselt number is always relative to that stagnant-film baseline, k ⁄ L — it is not a measure of temperature, flow speed, or heat transfer rate in absolute units, only of how far convection has out-performed conduction for the given geometry.
Where does the characteristic length come from?
It has to match whatever geometry produced the convective coefficient, h. For flow through a pipe, that is the pipe's internal diameter; for flow over a flat plate, it is the plate's length in the flow direction; for a sphere, its diameter. Using a length from a different correlation than the one that gave you h makes the ratio meaningless, even though the calculator will still divide the numbers without complaint.
Can this calculator find h for me from flow conditions?
No — it performs only the final ratio, Nu = hL ⁄ k. Getting h in the first place normally means computing Reynolds and Prandtl numbers for the flow and feeding them into an empirical correlation, such as the Dittus-Boelter equation for turbulent pipe flow. This instrument assumes h is already in hand, whether from such a correlation or from a direct measurement.
Why does Nu equal 1 mark the switch from conduction to convection?
Because at Nu = 1, the actual convective coefficient h exactly equals k ⁄ L, the coefficient a perfectly still film of the same thickness would give by conduction. Any Nu above 1 means the moving fluid is outperforming that stagnant baseline; Nu below 1 is unusual and generally signals an error in the inputs, since real convection rarely transfers heat worse than plain conduction would.
What happens to the Nusselt number if h doubles?
It doubles too, exactly, because the formula is linear in h. A coefficient of 25 W/(m²·K) with L = 0.5 m and k = 0.6 W/(m·K) gives Nu of about 20.83; raise h to 50 W/(m²·K) with everything else unchanged and Nu becomes about 41.67. Vigorous mixing, a higher flow velocity, or switching from natural to forced convection are the usual reasons h rises.
What happens to the Nusselt number if the length doubles?
It also doubles, for the same reason: the formula is linear in L as well as h. Stretch the characteristic length from 0.5 m to 1.0 m while holding h at 25 W/(m²·K) and k at 0.6 W/(m·K), and Nu rises from about 20.83 to 41.67 — the identical result that doubling h alone produced, since the two variables enter the ratio the same way.