Pump Cavitation vs. Air Binding on a Hot Condensate Transfer Pump
A condensate transfer pump takes suction from a vented condensate receiver at 210°F and discharges to a deaerator. The pump is a horizontal centrifugal unit rated 200 gpm at 120 ft head, driven at 3,550 rpm. Operators report the pump losing discharge pressure intermittently, a rattling noise, and vibration that peaks at about 1× running speed. The suction strainer differential pressure is 2 psi and stable. The receiver level is steady at 60%. The pump NPSH required is 6 ft; the available NPSH calculated from the receiver level and the suction piping is 9 ft at design flow. The pump has been in service two years without overhaul.
1. Distinguish between cavitation and air binding as the cause, and state what evidence would confirm each.
Read the worked answer
Both produce intermittent loss of discharge pressure, but the signatures differ.
Cavitation is a vapor-pressure phenomenon: the local pressure at the impeller eye drops below the vapor pressure of the 210°F condensate (about 14.7 psia at 212°F, so at 210°F the margin is thin). It produces a rattling or gravel-like noise, vibration that is broadband and often peaks at 1× and blade-pass frequency, and a discharge pressure that falls as flow increases. Confirming evidence: vibration analysis showing broadband energy and a 1× peak, a flow-versus-head curve that droops more than the factory curve, and - the decisive test - throttling suction or raising receiver level to increase NPSH available, which should reduce the noise.
Air binding is a gas-entrainment phenomenon: air or non-condensable gas accumulates at the impeller eye or in the suction line high point. It produces a softer, surging noise, discharge pressure that cycles slowly as the gas pocket forms and breaks, and vibration that is low-frequency and irregular rather than 1×. Confirming evidence: opening a vent at the pump casing high point releases air and restores pressure; the receiver vent may be plugged or the suction line may have a high point that traps gas.
The 1× peak and the rattling description point toward cavitation, but the 9 ft available versus 6 ft required margin is only 3 ft, and that margin can be eaten by a suction strainer loading, a partially closed suction valve, or a receiver level that drops during a transient. The stable 2 psi strainer dP argues against a fouled strainer, so the next check is the suction valve position and the actual receiver level during the pressure-loss events, not just at the moment of the reading.
2. What corrective actions would you take, in order, and what would you check before returning the pump to service?
Read the worked answer
Order of actions:
- Verify the suction path. Confirm the suction valve is fully open (a partially closed valve is a common cause of a 3 ft NPSH margin disappearing), check the receiver level trend during the events, and confirm the receiver vent is clear.
- Vent the pump casing at the high point to rule out air binding. If pressure restores and stays, the problem is gas, not cavitation.
- If cavitation is confirmed, raise the NPSH available: raise receiver level, reduce flow to lower NPSH required, or open the suction valve fully. Do not throttle discharge to fix a suction problem.
- Inspect the impeller at the next outage for cavitation damage - pitting on the suction side of the vanes, erosion at the eye. Two years without overhaul with intermittent cavitation is enough time for damage to accumulate.
- Check the pump alignment and bearings if vibration persists after the NPSH issue is corrected, because a 1× peak can also be misalignment or a worn bearing.
Before return to service:
- Confirm NPSH available exceeds required by a margin at the worst expected receiver level, not just at 60%.
- Verify the suction strainer differential pressure at design flow.
- Run the pump at design flow and confirm discharge pressure, noise, and vibration are back to baseline.
- Record the baseline vibration so the next event can be compared.
The exam point: cavitation and air binding look alike from the control room, and the cheap diagnostic is the casing vent. Fix the suction side, never the discharge.