Pressure Loss in Industrial Piping: Calculation, Tables and Impact on Pump NPSH

What is the difference between distributed and localized head loss?

The distributed loss occurs along the straight length through wall friction and depends on length, diameter, velocity, roughness and flow regime. The localized loss occurs at elbows, valves, tees, reducers, strainers, inlets and outlets and is represented by a K coefficient. The total loss is the sum of the two.

For industrial pump operators, head loss matters at two points: (1) at suction, it reduces the available NPSH — if the loss is too large, the pump cavitates; (2) at discharge, it increases the total head the pump must overcome. In a centrifugal pump, this shifts the operating point to the left on the curve, and the pump delivers less flow than designed.

In a positive displacement pump, such as the FBE and FBEI gear pumps, the flow stays practically the same and the discharge pressure rises instead — which is why the circuit needs a relief valve or another form of overpressure protection.

What is the distributed head loss formula?

Darcy-Weisbach computes Hf = f × (L/D) × v²/(2g), with Hf, L and D in meters, v in m/s and g = 9.81 m/s². The f in the formula is the Darcy friction factor, four times the Fanning factor used in part of the chemical engineering literature — mixing them up multiplies or divides the result by four.

The factor f depends on the Reynolds number and the relative roughness, and the equation holds for any Newtonian liquid when properties and regime are correctly specified.

The friction factor depends on the flow regime. In laminar flow — Reynolds below about 2,000, by the Hydraulic Institute criterion — f = 64/Re and the loss is proportional to viscosity and velocity; this is the common regime when pumping heavy oils and other viscous fluids with gear pumps.

In turbulent flow — Reynolds above about 4,000, the usual case for water in centrifugal pumps — f also depends on the relative roughness and comes from the Moody diagram or the Colebrook-White equation. Between 2,000 and 4,000 lies the critical zone, where the flow may be unstable and the friction factor is uncertain; avoid placing the design operating point there. The regime is set by the calculated Reynolds number, not by the type of pump.

For water and low-viscosity liquids, the established alternative in water supply and fire protection networks is the Hazen-Williams formula: Hf = 10.67 × L × Q^1.852 / (C^1.852 × D^4.87), with Q in m³/s and D and L in meters.

The C coefficient expresses the internal smoothness of the pipe and drops as the pipe ages: typical values are PVC ≈ 150 and galvanized iron ≈ 120, while cast iron, ≈ 130 when new, falls to 89–100 after about 20 years in service. In sprinkler networks, C is not a free choice: NFPA 13 sets the design value by material and system type — for black steel, 120 in wet systems and 100 in dry or preaction systems.

The formula's advantage is skipping the Reynolds number and the friction factor; its limitation is that it only applies to water, at ambient temperature and in turbulent flow. For oils and viscous fluids (FBE and FBEI pumps), use Darcy-Weisbach with the actual fluid viscosity at operating temperature.

The table below gives ready-to-use distributed head loss values per 100 m of straight pipe, calculated with Darcy-Weisbach and the Colebrook-White friction factor for water at about 20 °C (kinematic viscosity of 1.004 × 10⁻⁶ m²/s), in commercial carbon steel pipe (roughness ε = 0.045 mm) at the internal diameters shown. The values do not include fittings or design margin.

Note the effect of velocity: at 50 mm, doubling the flow from 10 to 20 m³/h multiplies the loss by 3.7 — slightly less than 4, because the friction factor drops as the Reynolds number rises.

When to use Darcy–Weisbach or Hazen–Williams?

Use Darcy-Weisbach for any Newtonian liquid, in any regime, with consistent viscosity, Reynolds number and roughness. Hazen-Williams is a practical approximation for water in turbulent flow and must not be transferred to oils or viscous fluids. In both cases, flow and diameter must be in the formula's units, and neither method removes the need to add the fitting losses.

How to calculate localized head loss?

Add the loss of each fitting via Hf = K × v²/(2g), using the velocity in the run where the fitting is installed. If several items share the same diameter, add their K coefficients before applying the formula. At reducers and expansions, check whether K was tabulated for the velocity on the smaller or the larger side — the smaller side is the usual reference. In critical suction lines, prefer K coefficients over equivalent length.

Localized loss occurs at elbows, tees, valves, reducers, expansions, strainers and at tank inlets and outlets. The coefficients in the table below are typical textbook values for turbulent flow (Purdue University minor-loss table; sudden expansion and contraction relations from the Clarkson University minor-loss notes).

They change with the exact geometry, the size, the connection type (flanged or threaded), the degree of opening and the manufacturer: the Hydraulic Institute itself records a ±50% spread among published values for flanged gate valves and +200% to −80% for check valves. Whenever available, use the fitting manufacturer's K or loss curve.

Equivalent length method: each fitting is converted into a run of straight pipe measured in diameters (L/D). The classic values from Crane Technical Paper 410 are: standard 90° elbow ≈ 30D, 45° elbow ≈ 16D, tee with flow through the run ≈ 20D and through the branch ≈ 60D, open gate valve ≈ 8D, open globe valve ≈ 340D and swing check valve ≈ 100D.

The conversion is K = f_T × (L/D), where f_T is the fully turbulent friction factor for that size — about 0.019 at 2″ and 0.015 at 6″. That is why equivalent length only holds in turbulent flow.

In laminar flow of viscous fluids, the real loss across fittings is much higher than tabulated: in critical NPSH designs, especially on the suction of FBE and FBEI gear pumps, calculate fitting by fitting with the K-coefficient method, corrected for low Reynolds numbers, such as Hooper's 2-K or Darby's 3-K methods.

How to calculate total head loss step by step?

First compute v = Q/A, using flow and internal diameter in compatible units. Then compute the distributed portion, add the localized portions and take Hf,total = Hf,distributed + Σ Hf,localized. Use the viscosity at the real operating temperature and report the result in meters of the pumped liquid.

The result is not yet the pump head. Total head adds the elevation difference between the suction and discharge levels, the difference between the required end pressure and the suction vessel pressure, converted to meters of liquid, and the total suction and discharge loss: H = Δz + (P2 − P1)/(ρ·g) + Hf,total, neglecting the kinetic energy difference between the vessels. With the design flow, H forms the (Q, H) duty point that goes to the pump curve.

For FBCN centrifugal pumps, check that point on the published curves at 1,750 rpm and 3,500 rpm (links below). If it falls within a model's operating range, send the point, the fluid and the temperature: engineering confirms the impeller, NPSH and power at the duty point.

How does suction head loss cause cavitation?

At suction, the loss enters as a direct subtraction: NPSHa = Pa ± Hz − Hf − Pv, with absolute and vapor pressures converted into meters of the liquid itself at operating temperature. Therefore, 1 m of additional loss reduces the available NPSH by 1 m and shrinks the margin over the required value.

Do not treat NPSHa = NPSHr as a no-cavitation condition: NPSH3, the basis of catalog NPSHr, is measured at the point where the pump head has already dropped 3% — that is, with cavitation under way. That is why ANSI/HI 9.6.1 expresses the margin as the NPSHa/NPSHr ratio and recommends values that vary with the type of application.

Numerical example, with hypothetical values only to show the order of magnitude: a centrifugal pump with an NPSHr of 3.2 m at the duty point, pumping water at 60 °C — atmospheric pressure at sea level ≈ 10.51 m and vapor pressure ≈ 2.07 m, both in meters of water at 60 °C. Pipe with a 65 mm internal diameter and a velocity of 2.5 m/s (≈ 30 m³/h; Re ≈ 343,000; f ≈ 0.019).

Good installation: flooded suction (Hz = +2.0 m), 5 m of pipe, sharp-edged inlet, 2 long-radius 90° elbows and 1 open gate valve (ΣK = 1.05). Distributed loss = 0.019 × (5/0.065) × (2.5²/19.62) = 0.47 m; localized = 1.05 × 0.319 = 0.33 m; Hf = 0.80 m; NPSHa = 10.51 + 2.0 − 0.80 − 2.07 = 9.64 m, a wide margin over the 3.2 m.

Poor installation: suction lift (Hz = −3.0 m), 15 m of pipe, sharp-edged inlet, 5 long-radius 90° elbows, a dirty Y-strainer (K = 8, assumed) and the gate valve (ΣK = 9.65). Distributed = 1.40 m; localized = 3.07 m; Hf = 4.47 m; NPSHa = 10.51 − 3.0 − 4.47 − 2.07 = 0.97 m — below the NPSHr, so the pump cavitates. Only the piping changed: same pump, same flow, same liquid.

How does viscosity change head loss?

In laminar flow, the friction factor is f = 64/Re, and the loss grows in proportion to viscosity for the same geometry and flow. Therefore, the calculation must use the viscosity at operating temperature and also check startup, when the fluid can be colder and far more viscous — often that is the condition that sizes the suction line.

In applications with FBE and FBEI gear pumps, the fluid is usually viscous — heavy oils, asphalt, resins — and suction flow tends toward the laminar regime. In that regime, at the same flow rate, the distributed loss falls with the fourth power of the diameter (Hagen-Poiseuille law): going from a 50 mm to a 65 mm suction line cuts that loss by about 65%.

That is why the suction pipe must never be smaller than the pump's suction nozzle and is often one or two sizes larger — a precaution that matters even more with viscous fluids. The FBE itself has a viscosity limit per size, according to the FB Bombas Engineering selection table (MTEC-01/01): 2,500 SSU from 1/8″ to 1/4″; 7,500 SSU from 3/8″ to 1″; 50,000 SSU at 1.1/2″ and 2″; and 100,000 SSU from 3″ to 6″.

From 1.1/2″ up, the table provides reduced speeds by pulley or gear reducer for the highest viscosities — down to 150 rpm from 3″ to 6″.

Another particularity: in many viscous fluids, viscosity depends strongly on temperature. Petroleum asphalt cement (CAP), for example, is semi-solid or solid at ambient temperature and only becomes fluid when heated, according to Petrobras technical information on asphalt. Head loss must be calculated for the worst condition, usually startup, when the fluid is coldest.

For this, the FBE has the CA version, with a heating chamber, available from 1.1/2″ to 6″: the heating medium circulates through the chamber and keeps the product fluid inside the pump. The suction line needs its own heating — steam, thermal oil or electric tracing — and thermal insulation.

Practical suction sizing rules

Practical rules to minimize suction head loss — the first four come from FB Bombas Engineering for the FBEI pumps (MAN001-10): (1) install the pump as close as possible to the suction source; (2) use as few fittings as possible on the suction line; (3) the suction pipe diameter must be at least equal to the pump's suction nozzle — never reduce before the pump; (4) use a strainer with the largest possible open area (the FBEI calls for a 40 to 80 mesh screen, depending on the model) and track clogging by differential pressure; (5) leave 5 to 10 diameters of straight pipe between the pump inlet and the first elbow, valve or tee; (6) on horizontal suction lines with a reducer, use an eccentric reducer with the flat side on top so that no air pocket forms; (7) on the suction side, use a full-bore isolation valve (gate or ball) and keep it fully open — throttling it adds loss exactly where it reduces NPSHa, and a globe valve (K ≈ 10) does not belong on the suction line; (8) prefer flooded suction whenever possible — each meter of positive head is one extra meter of NPSHa; (9) with viscous fluids on FBE pumps, respect the viscosity limit of the selected size: if the startup viscosity exceeds that limit, heat the fluid and the suction line before starting, or review the size and speed with engineering.

How to request pump selection from your calculation?

For a pump selection enquiry, send the calculation report, duty point (flow and head), fluid data (type, viscosity at operating and startup temperature, density), temperature and available NPSH at suction. If the calculation is not complete, include the piping layout (lengths, internal diameters, elevations and fittings), identifying measured and estimated values. To assess an FBCN, request the curve for the proposed configuration and identify the model, speed and impeller. FBCN 65-315 sheet CC-124.07 in the curves at 1,750 rpm (p.

25) shows these identifiers in its header and graphs; use the sheet corresponding to the equipment in your enquiry.