Pump Cavitation After a 10% Speed Increase: NPSH Available Versus NPSH Required
A poultry shed's water supply uses a centrifugal pump drawing from an open underground sump. The pump centreline sits 3.5 m above the lowest sump water level, and at the design duty the suction line loses 1.2 m of head to friction. Atmospheric pressure head is 10.3 m of water. Site measurements in May put the water temperature at 35 °C, at which the vapour pressure head is about 0.58 m. The nameplate gives NPSH required = 4.5 m at 1450 rpm. To meet higher summer demand, the motor is changed to a 1595 rpm unit and the same impeller is retained. Within a week the pump develops a crackling noise on the suction side, slight vibration and falling delivery. The fitter proposes partly closing the suction valve to "settle" the flow, and the owner wants to know whether the speed change is safe and what the correct remedy is.
1. Calculate the NPSH available at the original speed with the 35 °C water and state the margin over the required value.
Read the worked answer
Use the standard suction-side energy balance for an open sump:
NPSH_A = h_atm − h_lift − h_friction − h_vapour
NPSH_A = 10.3 − 3.5 − 1.2 − 0.58 = 5.02 m
Against NPSH_R = 4.5 m the margin is 5.02 − 4.5 = 0.52 m, roughly 12%.
The number is acceptable but thin. Two points matter for judgement: the vapour pressure head, not the friction loss, is the seasonal variable - the same installation in winter with cooler water would show a larger margin, so a winter commissioning test can hide a summer problem. And a margin under about 0.5–1.0 m leaves no room for a fouled strainer, a partly closed suction valve or the higher flow that follows any speed increase. The pump is not cavitation-proof; it is only just clear of cavitation.
2. Assess the 10% speed increase using the pump affinity relations and decide whether cavitation is now expected.
Read the worked answer
Flow scales with speed and friction loss scales with the square of flow, while NPSH_R scales with the square of speed:
- Flow: 1.10 × design flow, so suction friction = 1.2 × 1.10² = 1.2 × 1.21 = 1.45 m
- NPSH_R = 4.5 × 1.10² = 4.5 × 1.21 = 5.45 m
- NPSH_A = 10.3 − 3.5 − 1.45 − 0.58 = 4.77 m
NPSH_A (4.77 m) is 0.68 m short of NPSH_R (5.45 m), so the pump is now cavitating. That matches the symptoms: crackling noise from vapour bubbles collapsing on the impeller, vibration, and falling delivery as the head–flow curve degrades.
The common mistake is to hold NPSH_A constant. Raising speed does two things at once - it pushes the required value up with N² and drags the available value down because the higher flow increases suction friction. Both move in the wrong direction.
3. Evaluate the fitter's proposal to partly close the suction valve and recommend the correct remedy, ranking your options.
Read the worked answer
The fitter's proposal is wrong in direction. Partly closing the suction valve adds a deliberate loss on the suction side, so NPSH_A falls further - 1.0 m of extra loss would take the available value to about 3.77 m against a required 5.45 m, making cavitation worse, not better. A suction-side valve should be fully open in service and locked.
Ranked remedies:
- Best permanent fix: raise NPSH_A - lower the pump by at least 0.7 m, raise the sump water level, or increase the suction pipe diameter and remove fittings to cut friction. This restores the margin at any speed.
- Also sound: trim the impeller or restore the original speed/use a larger pump, i.e. match the duty rather than over-speed an existing unit.
- Acceptable stop-gap: throttle the discharge - flow falls, so both NPSH_R and suction friction drop and the available margin recovers; it wastes power and is not a solution.
- Reject: throttling the suction side, or ignoring the noise until the impeller erodes.