Absolute pipe roughness values by material

Values are compiled from the standard sources for this data: the Moody roughness classes, Crane TP-410, and the ASHRAE Handbook duct roughness classes for the ductwork entries. Where a material's published range is wide enough that any single number is a choice rather than a fact, the entry name says so and the value sits mid-range. Several entries deliberately carry a condition rather than only a material, because condition is usually the larger effect.

On provenance, honestly. The firm entries here are the classical Moody classes and the Crane values, which are reproduced consistently across the literature. The aged and corroded entries, the polyethylene and composite entries, and the lining entries are representative values inside a published range rather than a single traceable figure, and reasonable sources disagree about them. Where a lining or a proprietary product is involved, the supplier's own figure beats any table, including this one.

Relative roughness, and where roughness actually matters

Roughness never enters a friction calculation on its own. What matters is relative roughness, the absolute roughness divided by the internal diameter:

relative roughness = e / D

A roughness of 0.045 mm is smooth in a DN600 main and distinctly rough in a DN15 instrument line, so the same material sits in a different part of the Moody chart depending on the pipe it is in. Take the bore from the pipe schedule tables rather than the nominal size.

There is a second point that saves a great deal of pointless argument. In laminar flow the friction factor is 64 / Re and roughness does not appear at all, so for viscous oils, glycol at low temperature and small-bore laminar lines the roughness value you agonise over changes nothing. In the transitional and lower turbulent range roughness matters modestly. It is in fully rough turbulent flow, which is where most water and gas mains live, that roughness dominates and the friction factor stops depending on Reynolds number altogether. Knowing which regime you are in tells you how much the number below is worth arguing about.

Roughness is not surface finish

This is the most common misunderstanding about the numbers on this page, and it is worth being clear about.

Absolute roughness in the Colebrook-White sense is an equivalent sand-grain roughness. It comes from Nikuradse's experiments on pipes coated with sieved sand of known grain size, and a material's tabulated value is the sand-grain size that would produce the same friction as that material does. It is back-calculated from friction measurements, not measured with a profilometer.

So a mechanical surface-finish specification, an Ra in micrometres, is not the same quantity and does not convert cleanly. A pipe can have a fine Ra and a poor equivalent roughness because of joints, weld beads, ovality and waviness that a stylus tracing a short sample never sees. When someone offers you an Ra where a friction calculation wants an epsilon, ask for the hydraulic figure.

New pipe against the pipe you will actually have

Every value in the table above except those that name a condition is a value for new, clean pipe. Real pipe does not stay that way, and for metals in aggressive water the change over a design life can dwarf every other uncertainty in the calculation.

• Tuberculation and corrosion in unlined iron and steel carrying corrosive water can raise the effective roughness by an order of magnitude, and that is what the corroded entries in the table are for. Which of them applies to a given main is a judgement about water chemistry, age and history, not something a table can tell you.
• Scale from hard water narrows the bore as well as roughening it, and the bore reduction is often the larger effect because of the fifth-power dependence.
• Biofilm in warm, low-velocity water mains adds roughness without adding much scale.
• Linings do the reverse. A cement-mortar lining or an epoxy lining is applied precisely so the pipe does not do any of the above, which is why the lined entries sit two orders of magnitude below aged bare metal.

For an existing system where flow and pressure have both been measured, the correct move is not to choose a table value at all. Back-calculate the roughness that reproduces the measurement, and use that. A calibrated roughness for a specific main beats the best handbook value for its material.

Ductwork

The ductwork entries follow the ASHRAE roughness classes rather than the pipe convention, and they matter more than people expect because air systems run at low pressure and small absolute losses eat the fan's whole budget.

The flexible-duct entry deserves a warning. The tabulated value is for flexible duct that is fully extended. Compressed flexible duct, which is what most of it is once installed, has an effective resistance far worse than the table shows, and no roughness value will rescue a calculation that assumes an installation quality the job does not have.

Frequently asked questions

What is the absolute roughness of commercial steel pipe?

0.045 mm, which is the classical Moody value for clean commercial steel and the value in the table above. In metres, which is what the solver works in, that is 4.5e-5. Once the pipe has corroded the value climbs by a factor of three to forty depending on severity, which is why the library carries lightly, moderately and heavily corroded steel as three separate rows rather than one.

Is absolute roughness the same as relative roughness?

No. Absolute roughness is a length, usually quoted in mm, and it is a property of the material and its condition. Relative roughness is that length divided by the pipe's internal diameter, so it is dimensionless and depends on the pipe size. The Colebrook-White equation and the Moody chart both use the relative value.

Does roughness matter in laminar flow?

No. In fully laminar flow the Darcy friction factor is exactly 64 / Re, with no roughness term. This is a genuine result and not an approximation, and it means that for heavy oils, cold glycols and small-bore viscous lines the roughness value has no effect on the friction loss at all.

What roughness should I use for HDPE or PE100?

The library value is 0.003 mm, which reflects Australian water practice for PE100 mains. Published values for polyethylene range from about 0.0015 to 0.007 mm, which is a factor of five, but every value in that range is so smooth that the friction factor barely moves, so the choice rarely changes an answer. Manufacturer data for the specific pipe is available and is what to use if it matters.

How do I choose a roughness for old cast iron mains?

By condition and water chemistry, not by material. Unlined cast iron starts near 0.26 mm when new and reaches 1 to 3 mm and beyond when heavily tuberculated, a range wide enough that the material name alone tells you almost nothing. If you have any field measurement of head loss against flow, calibrate against it. If you have none, model the plausible range and see whether the answer to your actual question changes.

Which of these values does the solver use?

Whichever one is on the pipe. Choosing a material in the property panel writes its roughness onto the pipe and the value stays editable, so the library is a set of starting points rather than a constraint. The solver evaluates the roughness on each pipe through the Churchill (1977) correlation, which closely reproduces Colebrook-White in the turbulent range and reduces exactly to 64/Re in the laminar limit.

References

• Moody, L. F., Friction Factors for Pipe Flow, and the equivalent-roughness classes that follow from Nikuradse's sand-grain experiments and Colebrook's work on commercial pipes.
• Crane Co., Technical Paper No. 410: Flow of Fluids Through Valves, Fittings and Pipe. tp410.com
• ASHRAE Handbook, Fundamentals, for the duct roughness classes behind the ductwork entries.

Related

• Pipe schedule dimensions, for the internal diameter that turns absolute roughness into relative roughness.
• Hazen-Williams C factors, the empirical alternative that rolls roughness into one water-calibrated coefficient.
• Darcy friction factor calculator and the pipe flow and pressure drop calculator.
• Open the Studio to solve a network with these materials and see how much the choice actually moves the answer.