Suction Lift and NPSH Margin on a Summer Duty
A process pump in a textile plant lifts water from a sump whose surface is 4 m below the pump centreline. The suction line is 4 m of 80 mm inside-diameter pipe with one foot valve and strainer (K = 5.0) and two 90 degree bends (K = 0.75 each). Duty is 30 m³/h. The pump data sheet at that duty gives NPSHr = 4.5 m. Site data: atmospheric pressure 101.3 kPa, water density 996 kg/m³, Darcy friction factor 0.020. In April, with water at 30 °C and vapour pressure 4.24 kPa, the set runs quietly. In May the raw water reaches 40 °C, vapour pressure 7.38 kPa, and operators report rattling, falling discharge and impeller pitting within three weeks. The fitter proposes throttling the discharge valve slightly and replacing the suction line with 100 mm pipe. The plant manager wants the cheapest change that will survive every summer, and asks for numbers, not opinions.
1. Compute NPSHa in April and in May, compare with NPSHr, and state whether the April set-up was genuinely safe.
Read the worked answer
Q = 30/3600 = 8.333 × 10⁻³ m³/s; A = π(0.040)² = 5.027 × 10⁻³ m²; v = 1.658 m/s; v²/2g = 0.140 m.
Losses: friction fL/d = 0.020 × 4/0.080 = 1.00; fittings ΣK = 5.0 + 2 × 0.75 = 6.5; total coefficient 7.5, so hf = 7.5 × 0.140 = 1.05 m.
NPSHa = (p_atm − p_v)/(ρg) − H_s − hf. April (30 °C): (101.3 − 4.24) × 10³/(996 × 9.81) = 97 060/9770.8 = 9.93 m, so 9.93 − 4.00 − 1.05 = 4.88 m. Margin over NPSHr = 0.38 m, about 8 per cent. May (40 °C): (101.3 − 7.38) × 10³/9770.8 = 9.61 m, so 9.61 − 4.00 − 1.05 = 4.56 m. Margin = 0.06 m.
Judgement: April was not safe, only lucky. A 0.38 m margin is smaller than the uncertainty in the friction factor, and it disappears with a half-fouled strainer or a 0.3 m sump drawdown. May, at 0.06 m, is effectively running at NPSHr, which matches the rattling and the pitting.
2. Evaluate the fitter's two proposals - throttling the discharge and fitting a 100 mm suction - with numbers.
Read the worked answer
Throttling the discharge shifts the duty point to lower flow. Because NPSHr and suction loss both fall roughly with Q², a 15 per cent flow cut buys about 0.2 m of NPSHr and 0.3 m of hf, roughly 0.5 m of apparent margin, which is why the trick is popular. But it delivers less water to the process, wastes energy permanently, and cannot touch the 4 m static lift; the first fouled strainer takes the margin back. It is an emergency measure with a written limit, not a fix.
100 mm suction: v = 8.333 × 10⁻³/(π × 0.050²) = 1.061 m/s; v²/2g = 0.0574 m; coefficient = (0.020 × 4/0.100) + 6.5 = 0.8 + 6.5 = 7.3, so hf = 0.42 m. NPSHa in May = 9.61 − 4.00 − 0.42 = 5.19 m, margin 0.69 m.
That is a genuine gain and also cuts running loss, but it needs a new main, a reducer and a shutdown, and 0.69 m is still thin against a fouling strainer. Ranked on cost per metre of margin gained, it loses to attacking the static lift.
3. Recommend the cheapest change that survives every summer, with numbers, and state what to monitor.
Read the worked answer
Attack the dominant term. Static lift is 4.00 m while friction is only 1.05 m, so pipe size can never buy more than about 1 m, whereas lowering the pump buys 1.5 m for a modest plinth modification.
Lower the pump, or raise the sump level, by 1.5 m and keep the 80 mm line: H_s = 2.50 m, NPSHa in May = 9.61 − 2.50 − 1.05 = 6.06 m, a margin of 1.56 m - about 3.5 times the present May margin, and still over 1 m with a partly blocked strainer (+0.5 m) or a 0.3 m drawdown. Cost: civil work on the plinth and a short modified suction run, generally well below re-piping the suction main.
If the plinth cannot be lowered, the ranking is: vertical sump pump, which removes static lift entirely; then 100 mm suction with a low-NPSH impeller; then throttling as a temporary measure only.
Monitor suction pressure at the pump flange, since NPSHa = (p_s − p_v)/ρg + v²/2g; strainer differential pressure with a 0.3 m cleaning trigger; and daily water temperature in May and June. The 40 °C datum, not the 30 °C one, must be the design case.