Piping design and line sizing

Why it matters

Piping often costs as much as the equipment it connects, and every pump, compressor and control valve is sized from the pressure drop of the lines around it. Line sizing balances the capital cost of a bigger pipe against the pumping cost of a smaller one, while mechanical design makes sure the pipe wall, fittings and supports survive pressure, temperature and thermal expansion.

Key ideas

Line sizing by velocity. A practical first step is to choose a velocity from experience and calculate the diameter. Typical ranges (check your company or data-book guide): liquids in pump discharge lines about 1–3 m/s; pump suction lines lower, about 0.3–1.5 m/s, to protect NPSH; gravity lines lower still; gases and vapours about 15–30 m/s; steam somewhat higher. Very high velocities cause erosion, noise and water hammer; very low ones let solids settle and waste money on large pipe.

Economic (optimum) diameter. As diameter increases, the pipe and installation cost rises roughly linearly with diameter, while the pumping power falls roughly as D^(−5) at a fixed flow. The total annual cost has a minimum, the optimum diameter. Published correlations (for example in Peters and Timmerhaus or Coulson & Richardson Vol. 6) express this in terms of flow and density for carbon-steel pipe; they give a starting size which is then checked for velocity and pressure drop.

Pressure drop. For straight pipe use the Darcy–Weisbach equation with the Darcy friction factor f (which is 4 times the Fanning factor). In laminar flow (Re < 2100) f = 64/Re. In turbulent flow f depends on Re and relative roughness ε/D (Moody chart, Colebrook equation, or the explicit Swamee–Jain form). Commercial steel has ε ≈ 0.045 mm. Fittings and valves add losses written as K velocity heads (or as an equivalent length of pipe); K values come from data books. The line pressure drop, plus static head and equipment losses, fixes the pump head.

Nominal pipe size and schedule. Pipes are specified by nominal pipe size (NPS, in inches; the outside diameter is fixed for each size, e.g. NPS 4 has OD 114.3 mm and NPS 6 has OD 168.3 mm) and a schedule number that fixes the wall thickness, so a higher schedule means a thicker wall and a smaller bore. The inside diameter used in hydraulic calculations must come from the pipe tables, not from the nominal size. A rough guide: schedule number ≈ 1000 × P/S.

Wall thickness (ASME B31.3 style). The pressure-design thickness of straight pipe is t = P·D/(2·(S·E + P·Y)), using the outside diameter. Then add the corrosion allowance (and any thread or groove depth), and divide by (1 − mill tolerance) because seamless pipe may be up to 12.5 % thinner than nominal. Choose the first standard schedule whose wall is at least that.

Layout and flexibility. Hot lines expand by ΔL = α·L·ΔT. If both ends are fixed, the restrained expansion produces very large stresses and nozzle loads, so lines are routed with bends, expansion loops or bellows, guided and supported, and checked by flexibility analysis. Other considerations: slope for drainage, high-point vents and low-point drains, valve accessibility, avoiding pockets in vapour lines, and NPSH at pump suctions.

Material and standards. Pipe material follows the same logic as vessels (carbon steel by default; stainless, alloy, lined or plastic pipe for corrosive service). Process piping in India and internationally commonly follows ASME B31.3, with flanges to ASME B16.5 and pipe dimensions to ASME B36.10/B36.19 or equivalent IS standards.

Formulas

v = Q / A = 4·Q / (π·D²), D = √(4·Q / (π·v)) (line sizing) Re = ρ·v·D / μ ΔP_f = f·(L / D)·(ρ·v² / 2) (Darcy–Weisbach, Darcy f) f = 64 / Re (laminar) f = 0.25 / [log10(ε/(3.7·D) + 5.74/Re^0.9)]² (Swamee–Jain, turbulent) ΔP_fittings = ΣK·(ρ·v² / 2) t = P·D_o / (2·(S·E + P·Y)) (straight pipe, internal pressure; Y = 0.4 for ferritic and austenitic steels below about 480 °C) t_nom ≥ (t + CA) / (1 − 0.125) (12.5 % mill under-tolerance) ΔL = α·L·ΔT (thermal expansion)

• Q (m³/s); A (m²); D, D_o = inside and outside diameters (m or mm consistently); v (m/s); ρ (kg/m³); μ (Pa·s); L (m); ε = absolute roughness (m); K = loss coefficient (–, data book); P (MPa); S = allowable stress (MPa); E = quality factor (–); CA (mm); α = expansion coefficient (about 12 × 10⁻⁶ K⁻¹ for carbon steel).

Worked examples

Example 1 (standard): sizing a water line and its pressure drop Given: water 50 m³/h (ρ = 1000 kg/m³, μ = 1.0 × 10⁻³ Pa·s), target velocity about 2 m/s, line length 150 m, fittings ΣK = 8 (from data book), commercial steel ε = 0.045 mm. NPS 4 Schedule 40 has ID 102.3 mm.

1. Q = 50/3600 = 0.01389 m³/s; D = √(4 × 0.01389/(π × 2)) = 0.094 m → choose NPS 4 Sch 40 (ID 102.3 mm).
2. Actual velocity: v = 0.01389 / (π/4 × 0.1023²) = 1.69 m/s.
3. Re = 1000 × 1.69 × 0.1023 / 10⁻³ = 1.73 × 10⁵ (turbulent).
4. f = 0.25 / [log10(0.045/(3.7 × 102.3) + 5.74/(1.73 × 10⁵)^0.9)]² = 0.0189.
5. Pipe: ΔP = 0.0189 × (150/0.1023) × (1000 × 1.69²/2) = 39.5 kPa; fittings: 8 × 1000 × 1.69²/2 = 11.4 kPa.
6. Total friction loss ≈ 50.9 kPa, about 5.2 m of water head, to be added to the static head for pump sizing.

Example 2 (GATE level): wall thickness and schedule, plus expansion Given: NPS 6 carbon-steel line, OD 168.3 mm, design pressure 4.0 MPa, S = 138 MPa, E = 1.0 (seamless), Y = 0.4, CA = 1.5 mm, mill tolerance 12.5 %. NPS 6 walls: Sch 10 = 3.40 mm, Sch 40 = 7.11 mm. The 50 m line runs from 20 °C to 200 °C, α = 12 × 10⁻⁶ K⁻¹.

1. t = 4.0 × 168.3 / (2 × (138 × 1.0 + 4.0 × 0.4)) = 673.2 / 279.2 = 2.41 mm.
2. Add CA: 2.41 + 1.5 = 3.91 mm; allow for under-tolerance: 3.91 / 0.875 = 4.47 mm.
3. Sch 10 (3.40 mm) is too thin; Sch 40 (7.11 mm) is chosen. Check: 1000 × P/S = 1000 × 4/138 ≈ 29, consistent with Sch 40.
4. Expansion: ΔL = 12 × 10⁻⁶ × 50 × 180 = 0.108 m = 108 mm, which must be absorbed by bends, a loop or bellows, not by the equipment nozzles.

Common mistakes

• Using the nominal size (e.g. 4 inch = 101.6 mm) as the inside diameter.
• Mixing Fanning and Darcy friction factors (a factor of 4 error).
• Forgetting fittings and valves, which can be a large share of the loss in short plant lines.
• Using inside diameter in the B31.3 thickness formula, which is written with outside diameter.
• Forgetting the 12.5 % mill tolerance and corrosion allowance when picking a schedule.
• Sizing pump suction lines at discharge velocities, starving the pump of NPSH.

For GATE CH

Expect numericals on velocity-based line sizing, Reynolds number, Darcy–Weisbach pressure drop with a given friction factor, fitting losses with K values, and conceptual questions on economic pipe diameter, schedule numbers and thermal expansion. Practise unit conversions from m³/h and keeping f consistent.

Quick check

1. What diameter carries 0.02 m³/s at 2.5 m/s?
2. Darcy f is 0.02. What is the Fanning factor?
3. Why are suction lines sized for lower velocity than discharge lines?
4. By how much does a 30 m steel line lengthen when heated by 150 K?

Answers: 1. √(4 × 0.02/(π × 2.5)) = 0.101 m. 2. 0.005. 3. To keep friction loss low and preserve NPSH available. 4. 12 × 10⁻⁶ × 30 × 150 = 0.054 m = 54 mm.