Condensate Pump Cavitation After a Heat-Recovery Retrofit
An atmospheric condensate receiver at a food plant serves a 200 gpm centrifugal pump. The pump was selected three years ago, when condensate entered the receiver at 180°F; a heat-recovery project now returns it at 205°F. The receiver vents to atmosphere, its normal liquid level is 4 ft above the pump centerline, and the 3 in. suction line (25 ft equivalent length, clean strainer) drops 3.5 ft of head at 200 gpm. Published pump NPSHr at 200 gpm is 9 ft, and the original selection left roughly 8 ft of margin. Since heat-recovery startup, operators report gravelly noise, 10–15% oscillation on the magnetic flowmeter, and two mechanical-seal failures in four months. The low-level alarm point sits 18 in. below normal level, and the pump trips on low flow during weekend washdowns. Maintenance has proposed adding a second strainer and buying a larger pump to "push through" the problem.
1. Compute NPSHa for the original 180°F condition and for today's 205°F condition, and determine whether cavitation explains the symptoms.
Read the worked answer
Use NPSHa = (p_atm − p_v)(2.31 ft/psi) + z − h_f, with absolute pressures at the pump suction flange.
Original design, 180°F: p_v ≈ 7.5 psia, so (14.7 − 7.5)(2.31) = 16.6 ft. Add 4 ft of static submergence and subtract 3.5 ft of suction friction: NPSHa = 17.1 ft. Against NPSHr = 9 ft that is 8.1 ft of margin, matching the original selection. The pump was correctly applied at the time.
Today, 205°F: p_v ≈ 12.7 psia, so (14.7 − 12.7)(2.31) = 4.6 ft. NPSHa = 4.6 + 4 − 3.5 = 5.1 ft. Margin = 5.1 − 9 = −3.9 ft; the pump is being run well below its required NPSH. At the low-level alarm, remove another 1.5 ft and NPSHa falls to about 3.6 ft.
The diagnosis is consistent with every symptom: near-saturated liquid at the suction, head loss that grows with flow, and a low-level condition that makes it worse. Gravelly noise, flow oscillation, and repeated seal failures are classic cavitation signatures, and the weekend low-flow trips occur when the receiver level is lowest. Nothing in the data suggests a mechanical or electrical fault.
2. Rank the practical fixes quantitatively. Evaluate: (a) adding a strainer, (b) a bigger pump, (c) raising the receiver 4 ft, (d) upsizing the suction line to 4 in., (e) cooling the condensate to 180°F, (f) a low-NPSH pump.
Read the worked answer
Compare each option against the required 9 ft plus a 2–3 ft margin, i.e. a target of about 11–12 ft NPSHa at 205°F.
- (a) Adding a strainer - reject. It removes margin; a dirty strainer can cost several feet exactly when flow is highest.
- (b) A bigger pump - reject. Higher flow raises NPSHr and suction friction simultaneously, so the margin gets worse, not better.
- (c) Raise receiver 4 ft: NPSHa = 5.1 + 4 = 9.1 ft, which is zero margin. Only viable combined with other changes.
- (d) 4 in. suction: friction falls from 3.5 ft to roughly 0.9 ft at 200 gpm, giving 5.1 + 2.6 = 7.7 ft. Still short of NPSHr.
- (e) Cool condensate to 180°F: restores p_v to 7.5 psia and NPSHa to about 17 ft, the largest single gain, but it needs a roughly 2.5 MMBtu/h exchanger and a useful sink for the heat.
- (f) Low-NPSH pump (NPSHr ≈ 3 ft): gives 5.1 − 3 = 2.1 ft margin - workable, and cheapest if a suitable hydraulically selected pump exists.
Best combination is (f) plus (d) for about 4.8 ft of margin, or (e) if the heat can be recovered. Also raise the low-level alarm and trip points and stop the washdown trips by adding a low-flow recirculation line. Do not throttle the discharge to fix this; discharge throttling changes neither NPSHa nor NPSHr.
3. What measurements would confirm cavitation rather than a mechanical fault, and what operating limits would you set until the fix is installed?
Read the worked answer
Confirming measurements:
- Vibration spectrum at the pump bearing housing. Cavitation shows broadband high-frequency energy (2–10 kHz) plus a peak at vane-pass frequency (for a 5-vane impeller at 3,500 rpm, 5 × 58.3 Hz ≈ 292 Hz). Misalignment or bearing defects instead show 1× running speed or bearing defect frequencies.
- NPSH test: throttle the suction valve in small steps while logging flow and noise. If available NPSH is the limit, flow will begin to break away and noise will rise at a suction head close to the calculated 5 ft; a mechanically healthy pump keeps its head down to NPSHr.
- Flow and head at constant speed compared to the factory curve. Cavitating pumps fall below the curve at high flow and recover at low flow.
- Seal inspection: cavitation damage tends to appear on the suction side of the seal faces and the impeller inlet, not evenly around the shaft.
Interim limits: hold receiver temperature at or below about 195°F (p_v ≈ 10.6 psia, NPSHa ≈ 9.5 + 4 − 3.5 = 10 ft, barely 1 ft margin), keep level at or above normal, cap flow near 180 gpm to limit suction friction, and keep the discharge recirculation line open during washdowns. Document this as a temporary restriction, not a solution.