Why we bother with minor losses
Pipe friction is the bulk of headloss in a long-run pipeline. Fitting losses (elbows, valves, tees, expansions, contractions) are individually small but collectively significant — especially in:
- Short pipe runs with many fittings (pump room piping, plant interiors)
- Throttled or partially-open valves (where K shoots up dramatically)
- Suction-side piping (where NPSHr margin is tight)
In a 100-ft pump room piping run with 12 elbows, 4 valves, a reducer, and a strainer, minor losses can easily exceed friction. "Minor" refers to the dimension of each individual loss, not the cumulative effect.
The K-factor formula
h_L = K · V² / (2g)
Where:
- h_L = headloss across the fitting (ft of fluid)
- K = dimensionless resistance coefficient
- V = mean velocity in the fitting (ft/s)
- g = 32.2 ft/s²
K is roughly constant for fully-turbulent flow (Re > ~10^4 for most fittings). At lower Re, K rises significantly; most plant-piping designs run in the turbulent zone where the constant-K approximation works.
Typical K-values (turbulent flow, full-port, full-open)
| Fitting | K | |---|---| | 90° standard elbow (threaded) | 0.9 | | 90° long-radius elbow (flanged) | 0.30 | | 45° elbow | 0.40 | | Tee, flow through run | 0.6 | | Tee, flow through branch | 1.8 | | Gate valve, full open | 0.15 | | Gate valve, 50% open | 4.5 | | Gate valve, 25% open | 24 | | Globe valve, full open | 6.0 | | Ball valve, full port, full open | 0.05 | | Butterfly valve, full open | 0.86 | | Butterfly valve, 70° open | 1.4 | | Swing check valve | 2.0 | | Lift check valve | 12 | | Strainer (mesh, clean) | 1.5 | | Strainer (mesh, dirty) | 4-10 (varies wildly) | | Pipe entrance, sharp-edged | 0.5 | | Pipe entrance, well-rounded | 0.05 | | Pipe exit (to tank) | 1.0 | | Sudden expansion (small to big, area ratio = 0.25) | 0.56 | | Sudden contraction (big to small, area ratio = 0.25) | 0.39 |
Values vary 20-30% between references (Crane TP-410 vs. Hydraulic Institute vs. ASHRAE). Use one source consistently rather than mixing.
The "equivalent length" alternative
Some texts express fitting losses as equivalent length of straight pipe:
h_L = f · (L_eq / D) · V² / (2g)
So:
K = f · (L_eq / D)
The advantage: you can add equivalent lengths to actual pipe length and run one Hazen-Williams or Darcy-Weisbach calculation. The disadvantage: equivalent length depends on f, which depends on Re — so it changes with flow. K is closer to a constant.
For modern design, stick with K-factors. Equivalent length is a holdover from slide-rule days.
Sudden expansions and contractions (Borda-Carnot)
A sudden expansion has a derivable K, not empirical:
K_expansion = (1 − A₁/A₂)²
For full-area expansion to a tank (A₂ → ∞): K = 1.0. That's the K for pipe-exit-to-tank.
Sudden contraction is empirical and approximately:
K_contraction ≈ 0.5 · (1 − A₂/A₁)
The asymmetry — expansion losses are larger than contraction losses for the same area ratio — reflects the difference between expanding flow (with separation, recirculation, dissipation) and contracting flow (which accelerates smoothly).
When throttling, K dominates
Closing a valve part-way is the most dramatic K change in normal operation. A gate valve dropping from full-open (K = 0.15) to 25% open (K = 24) becomes the single largest loss in any piping system that contains one.
Two practical consequences:
1. Don't size a system at full-open valve K. If you intend to throttle for flow control, design with the throttled K so the pump curve crosses the system curve at the throttled flow. 2. Use the right valve type for throttling. Globe valves are designed for it; gate valves are not. Throttling with a gate valve causes cavitation damage to the seat surfaces. If the operating plan calls for throttling, spec a globe or control valve.
The strainer trap
Strainer K is usually quoted at clean condition. In service, debris loads the screen and K rises rapidly. A strainer at K = 1.5 clean may climb to K = 5-10 within months in untreated water service.
The design move: include a clean K in the design point and a dirty-strainer K equal to 3-4× the clean value as a check case. Verify the pump still operates within its AOR and the system still meets flow requirements at the dirty-strainer condition. If not, plan strainer maintenance into the operating procedure.
Suction-side K and NPSH
The K-factor matters most on the suction side because every foot of head lost there comes off NPSHa. A poorly-laid-out suction piping run with three elbows and a partially-clogged strainer can easily eat 8-10 ft of NPSHa — turning a comfortable pump selection into a cavitating one.
Rules:
- Minimize fittings on the suction side. Aim for 2 elbows or fewer between the source and the pump suction.
- Use long-radius elbows (K ~ 0.30 vs. K ~ 0.9 for standard).
- Eccentric reducers, flat-top. When transitioning a horizontal suction pipe down to a smaller pump suction, use an eccentric reducer with the flat side up. Concentric reducers trap air at the top.
- No bushings. Bushings (a small reducer threaded into a larger fitting) have K values 4-5× the published reducer K because of the abrupt step.
Putting it together — worked example
Pump room piping, 6" SCH 40, 100 gpm water.
V = 100 / (2.448 · 6²) = 1.13 ft/s (roughly, for nominal sizing)
Actually for 6" SCH 40 ID = 6.065 in, A = 0.2006 ft²:
V = (100 gpm) · (0.002228 ft³/s/gpm) / 0.2006 ft²
= 1.11 ft/s
V²/(2g) = 1.11² / 64.4 = 0.019 ft of velocity head.
Now sum the K's:
| Fitting | Quantity | K (each) | ΣK | |---|---|---|---| | Long-radius 90° elbow | 4 | 0.30 | 1.20 | | Ball valve (full open) | 2 | 0.05 | 0.10 | | Tee, flow-through-run | 1 | 0.6 | 0.60 | | Strainer (clean) | 1 | 1.5 | 1.50 | | Suction entrance, well-rounded | 1 | 0.05 | 0.05 | | Discharge to header (sudden expansion to tank) | 1 | 1.0 | 1.00 | | Total | | | 4.45 |
Headloss:
h_L = 4.45 · 0.019 = 0.085 ft
About one inch of head from minor losses in this segment.
Now suppose the duty doubles to 200 gpm. Velocity doubles to 2.22 ft/s; velocity head quadruples to 0.077 ft.
h_L = 4.45 · 0.077 = 0.34 ft
About 4 inches. Quadruples when flow doubles — same Q² scaling as Hazen-Williams pipe friction. Total system curve impact scales cleanly even if the fitting count doesn't change.
When to neglect minor losses
The classical engineering rule: if minor-loss headloss < 5% of pipe-friction headloss, treat as zero. In a 5,000-ft force main with maybe 6 elbows and 4 valves, minor losses are typically 1-2% of pipe friction — negligible.
In a pump room with 60 ft of pipe and 25 fittings, minor losses can be 50-100% of pipe friction — definitely not negligible.
The 5% rule is a screen, not a license to skip the calculation. It takes 30 seconds to compute the sum-of-K and compare. Do that, then decide.
How the calculator handles it
The Headloss Calculator's fittings panel lets you specify quantity-by-fitting-type from a dropdown of standard fittings with published K-values. The calculator sums ΣK and applies h_L = ΣK · V²/(2g) at every velocity along the system curve.
Add fittings → see system curve climb → see operating point move → see whether the pump still fits.
Run a minor-loss calculation →
References
- Crane Co. *TP-410: Flow of Fluids Through Valves, Fittings, and Pipe.* (Standard reference, K-values + equivalent lengths.)
- Idelchik, I. E. *Handbook of Hydraulic Resistance,* 3rd ed. (Encyclopedic — fittings, transitions, manifolds.)
- Hydraulic Institute. *ANSI/HI 9.6.7 — Rotodynamic Pumps — Guideline for Effects of Pumping on Liquid Viscous Pumping.*
- ASHRAE Handbook — *Pipe sizing chapter.*