Suction-Starved Cooling-Water Pump: NPSH Margin or a Throttling Orifice?
A municipal cooling-water pump draws from an open sump and delivers 120 m³/h at 45 m head when running at 1470 r/min. The suction line is 150 mm bore, 12 m long, with four long-radius elbows and a basket strainer; the pump centreline sits 3.5 m above sump minimum level. Summer water temperature reaches 35 °C (vapour pressure 5.6 kPa). Site readings at full flow: suction-flange gauge −45 kPa, discharge 3.2 bar g, motor current 96% of rated. Operators report a gravelly noise and slight vibration that appear only above about 110 m³/h and disappear when the discharge valve is throttled. The maintenance planner proposes fitting a permanent throttling orifice and closing the issue. The strainer was last cleaned 14 months ago; differential pressure across it is not instrumented. The pump vendor's NPSHr is 3.8 m at 120 m³/h and 5.0 m at 132 m³/h.
1. Calculate the available NPSH at the measured 120 m³/h condition and at approximately 110% flow. State your assumptions and judge whether cavitation is plausible at the higher flow.
Read the worked answer
At the suction flange, absolute pressure = 101.3 − 45 = 56.3 kPa (assuming sea-level atmospheric pressure and a gauge in calibration).
NPSHa = (p_s − p_v)/(ρg) + v²/(2g) + z, with the flange taken as the reference (z = 0):
- Pressure head: (56.3 − 5.6)/9.81 = 5.17 m
- Flow area: A = π(0.150)²/4 = 0.01767 m²; Q = 120/3600 = 0.0333 m³/s; v = 1.89 m/s
- Velocity head: 1.89²/(2 × 9.81) = 0.18 m
- NPSHa ≈ 5.35 m at 120 m³/h, against NPSHr 3.8 m → margin about 1.55 m
At 132 m³/h the velocity rises to about 2.07 m/s (velocity head 0.22 m), and suction-line losses rise roughly with Q². The strainer and pipe losses are already embedded in the measured flange pressure, so extrapolating from 120 to 132 m³/h adds perhaps 0.3–0.5 m of loss. NPSHa falls to roughly 4.9–5.0 m while NPSHr rises to 5.0 m - essentially zero margin.
That matches the observed threshold: the noise begins only when the operating point moves far enough up the curve. The calculation is sensitive to gauge calibration and sump level, so both should be verified before the result is treated as exact.
2. Explain the physical meaning of the gravelly noise and vibration. Is permanent discharge throttling an acceptable engineering fix?
Read the worked answer
The noise is characteristic of cavitation: local static pressure in the impeller eye or suction passages falls below vapour pressure, bubbles form and then collapse as the fluid moves to a higher-pressure region. Collapse produces the crackling or gravelly sound, vibration, and eventually pitting of the impeller and casing. It appears at high flow because the pump's required NPSH rises with flow while the system's available NPSH falls as suction losses increase.
Throttling the discharge reduces flow, which moves operation back down the curve: NPSHr falls, suction loss falls, and the noise stops. It therefore treats the symptom, but it is not an acceptable permanent fix for several reasons:
- It wastes energy by creating a deliberate pressure drop; the pump still runs at essentially full speed.
- It hides the real cause, which is likely a fouled strainer, an undersized or loss-heavy suction line, or summer water temperature combined with low sump level.
- It reduces the plant's available capacity and may simply move the cavitation threshold to a condition that the operators later need.
- Cavitation damage already incurred is not reversed, and continued operation near the threshold can still cause erosion.
Throttling is defensible only as a temporary operating restriction while the suction problem is corrected, and the restriction should be documented as such.
3. Recommend a corrective package, including instrumentation and operating changes, and comment on the energy and reliability consequences.
Read the worked answer
A sensible sequence is:
- Verify the measurement: confirm gauge calibration, sump level, and water temperature. Recalculate NPSHa with the actual atmospheric pressure at site altitude.
- Clean or replace the strainer and install a differential-pressure transmitter with an alarm. A 14-month cleaning interval with no DP indication is the most likely immediate cause; the strainer loss appears directly in the flange pressure.
- Check the suction line: 1.9 m/s is not excessive, but four elbows and a long run add loss. Replace elbows with long-radius fittings, remove unnecessary valves, and consider increasing the suction bore if the margin remains thin at maximum summer temperature.
- Restore margin at the source: keep sump level high, consider a suction cooling or blending option if the water temperature cannot be reduced, and confirm the pump is not oversized for the actual duty.
- If the duty varies, fit a variable-speed drive and operate near the best-efficiency point. Reducing speed lowers both flow and NPSHr, so the margin improves without throttling loss.
- Until corrections are complete, impose a temporary maximum-flow limit and trend vibration and noise, not just pressure.
Energy-wise, a VFD on a predominantly friction system can cut power roughly with the cube of speed; even a modest speed reduction is far cheaper than a permanent orifice. Reliability-wise, removing cavitation protects the impeller, mechanical seal, and bearings, and instrumenting the strainer converts an invisible fouling problem into a planned maintenance task.