The Moody Chart

Why the Moody chart matters

Every time a fluid moves through a pipe, the wall drags on it and dissipates mechanical energy as heat. That loss sets the pump you must buy, the pressure your city water reaches, the fan power a data center burns to move air, and the throughput of an oil pipeline hundreds of kilometres long. The physics is captured by a single dimensionless number — the Darcy friction factor f — but f itself depends on the flow in a way no closed-form equation handled cleanly before 1944. The Moody chart solved that: it puts the entire behaviour of pipe friction, from creeping laminar flow to raging fully-rough turbulence, on one page you can read with a ruler.

• Pump and fan sizing. The head loss from Darcy-Weisbach sets the operating point on a pump curve.
• Pipe-diameter selection. A slightly larger diameter can slash pumping cost because loss scales with 1/D⁵.
• Water distribution. Municipal and building plumbing networks are balanced against these losses.
• Oil and gas pipelines. Long-haul lines live in the fully-rough regime where roughness dominates.
• HVAC ductwork. Air handling uses the same chart with the hydraulic diameter of rectangular ducts.
• Process plants. Chemical, food, and pharma piping is sized loop by loop against friction loss.

How to read it, step by step

The chart has three inputs and one output. You compute two dimensionless groups, find your operating point, and read f.

1. Compute the Reynolds number. Re = ρVD/μ = VD/ν, where ρ is density (kg/m³), V is the mean (bulk) velocity (m/s), D is inside diameter (m), μ is dynamic viscosity (Pa·s), and ν = μ/ρ is kinematic viscosity (m²/s). This locates you on the horizontal axis, which spans roughly Re = 10³ to 10⁸.
2. Compute the relative roughness. ε/D, where ε is the equivalent sand-grain roughness of the wall material (m) and D is the inside diameter (m). This picks which curve you follow — each labelled curve on the right side of the chart is one value of ε/D.
3. Choose the regime. If Re < 2300 you are on the laminar line and roughness is irrelevant. If Re > 4000 you are on the turbulent curves. Between them is the transition band, which the chart deliberately leaves ambiguous.
4. Read the friction factor. Drop a vertical line from your Re, find where it crosses your ε/D curve, and read f off the vertical axis (roughly 0.008 to 0.1 in practice).
5. Feed Darcy-Weisbach. Put f into h_f = f (L/D)(V²/2g) to get head loss in metres of fluid, or Δp = f (L/D)(ρV²/2) for pressure drop in pascals.

The governing turbulent curve is the Colebrook-White equation, which the whole right half of the chart is simply a plot of:

1/√f = −2·log₁₀( (ε/D)/3.7 + 2.51 / (Re·√f) )

Here f is the Darcy friction factor (dimensionless), Re is the Reynolds number (dimensionless), and ε/D is the relative roughness (dimensionless). Because f sits inside the logarithm on the right and outside it on the left, the equation is implicit — you cannot solve for f algebraically. Two limits fall out of it. When ε/D → 0 the roughness term vanishes and you recover the smooth-pipe Prandtl law, 1/√f = 2·log₁₀(Re·√f) − 0.8. When Re → ∞ the viscous term 2.51/(Re√f) vanishes and you recover the fully-rough von Kármán law, 1/√f = −2·log₁₀( ε/(3.7·D) ), which has no Re in it at all — that is why those curves go flat.

Because iterating Colebrook by hand is tedious, engineers often use the explicit Swamee-Jain approximation, valid for 10⁻⁶ ≤ ε/D ≤ 10⁻² and 5×10³ ≤ Re ≤ 10⁸:

f = 0.25 / [ log₁₀( (ε/D)/3.7 + 5.74/Re^0.9 ) ]²

It reproduces the Colebrook chart to within about ±1 percent — well inside the ±5 to ±10 percent uncertainty of the roughness value ε itself, which is the real weak link in any of these calculations.

The four regions of the chart

Reading the Moody chart is really about knowing which of four regions your operating point lands in, because the physics — and the sensitivity of f to your inputs — is different in each.

Region	Reynolds range	Depends on	Governing law
Laminar	Re < 2300	Re only	f = 64/Re
Transition (critical)	2300 – 4000	unstable / indeterminate	none reliable
Smooth-pipe turbulent	> 4000, low ε/D	Re only	Prandtl: 1/√f = 2 log₁₀(Re√f) − 0.8
Fully-rough turbulent	high Re, high ε/D	ε/D only	von Kármán: 1/√f = −2 log₁₀(ε/3.7D)

The two turbulent regions are connected by the general Colebrook curve, and the dashed line running diagonally across the chart marks the boundary where roughness effects start to dominate over viscous effects — the entry into fully-rough behaviour. A pipe that is "smooth" at low velocity can become "rough" simply by pumping faster, because raising Re thins the viscous sublayer until the roughness peaks poke through it.

Typical equivalent roughness values ε that set which curve a real pipe rides on:

Material	ε (mm)	ε/D at D = 100 mm
Drawn tubing (glass, brass, copper)	0.0015	0.000015
Commercial / welded steel	0.045	0.00045
Galvanized iron	0.15	0.0015
Cast iron (asphalted)	0.12	0.0012
Cast iron (uncoated)	0.26	0.0026
Concrete	0.3 – 3.0	0.003 – 0.03
Riveted steel	0.9 – 9.0	0.009 – 0.09

Common misconceptions and failure modes

• Darcy versus Fanning. The Moody chart uses the Darcy friction factor (f = 64/Re laminar). The Fanning factor, common in chemical engineering, is exactly one quarter of it (f_F = 16/Re). Mixing them gives a 4× error in head loss.
• Roughness matters in laminar flow. It does not. Below Re ≈ 2300 the viscous sublayer fills the pipe; f = 64/Re regardless of wall texture. Reading a rough-pipe curve there is wrong.
• New-pipe roughness lasts. Corrosion, scaling, and biofilm can raise ε by an order of magnitude over decades. Old cast iron can behave like riveted steel, so design roughness is a life-cycle choice, not a catalog value.
• Designing in the transition band. Between Re 2300 and 4000, f is genuinely indeterminate and can double for a whisker of change. Robust designs stay out of it or take the conservative turbulent value.
• Confusing ε with a physical bump height. ε is an equivalent sand-grain roughness calibrated to Nikuradse's experiments, not a measured surface profile. Two walls with the same Ra can have different ε.
• Ignoring temperature. Viscosity ν changes Re strongly. Hot water has roughly one-third the viscosity of cold water, shifting Re — and therefore f — noticeably.
• Non-circular ducts. The chart is for round pipes. For ducts, substitute the hydraulic diameter D_h = 4A/P (four times area over wetted perimeter); it is an approximation, best in turbulent flow.

Worked example: head loss in a water main

Water at 20 °C flows at V = 2 m/s through a commercial-steel pipe of inside diameter D = 100 mm over a length L = 500 m. Find the head loss and pressure drop.

1. Fluid properties. ρ = 998 kg/m³, ν = 1.004×10⁻⁶ m²/s at 20 °C.
2. Reynolds number. Re = VD/ν = (2)(0.10)/(1.004×10⁻⁶) ≈ 1.99×10⁵. That is well above 4000, so the flow is turbulent.
3. Relative roughness. Commercial steel ε ≈ 0.045 mm, so ε/D = 0.045/100 = 4.5×10⁻⁴.
4. Friction factor. Entering the Moody chart at Re ≈ 2×10⁵ on the ε/D = 0.00045 curve gives f ≈ 0.019. The Swamee-Jain formula confirms f = 0.25/[log₁₀(0.00045/3.7 + 5.74/(1.99×10⁵)^0.9)]² ≈ 0.0187.
5. Head loss. h_f = f (L/D)(V²/2g) = 0.0187 × (500/0.10) × (2²)/(2×9.81) = 0.0187 × 5000 × 0.2039 ≈ 19.1 m of water.
6. Pressure drop. Δp = ρ g h_f = 998 × 9.81 × 19.1 ≈ 1.87×10⁵ Pa ≈ 1.87 bar.

Notice how brutally loss scales with the variables: because h_f ∝ f L V²/D and, for a fixed flow rate Q, V ∝ 1/D², head loss goes roughly as 1/D⁵. Bumping this pipe from 100 mm to 125 mm inside diameter would cut the head loss by more than a factor of three — often cheaper over a plant's life than the extra pumping power. That trade-off, read straight off the Moody chart, is the daily bread of hydraulic design.