NPSH Checks for Chilled Water Pumps Before Diagnosing Cavitation
AI-generated editorial illustration; not a verified installation or construction detail.
Start with a suction pressure calculation
A rattling pump deserves attention, but sound alone does not identify cavitation. A useful first screen combines measured suction pressure, local atmospheric pressure, fluid temperature, flow and the manufacturer’s pump data.
Net positive suction head available, or NPSHA, expresses how far the liquid’s total suction head is above its vapor-pressure head. It belongs to the installation and operating condition. NPSH required, or NPSHR, comes from the pump manufacturer for a particular flow, speed and liquid. A positive gauge-pressure reading can still leave insufficient margin for the selected pump.
Keep gauge and absolute pressure separate
A gauge reading uses local atmospheric pressure as its zero. Vapor pressure uses a vacuum reference. Convert them to the same reference before subtracting: absolute suction pressure = gauge suction pressure + local atmospheric pressure.
For a suction-flange measurement at the manufacturer’s NPSH datum, a useful form is NPSHA = (p_suction,gauge + p_atmosphere − p_vapor)/(rho × g) + v²/(2g). Use pascals for pressure, kg/m³ for density and m/s for velocity to obtain metres of the pumped liquid. The Hydraulic Institute pump principles reference defines NPSHA using total absolute suction head minus vapor-pressure head.
If the tap is remote, account for elevation, velocity changes and losses between the tap and the pump reference section. A reading upstream of a suction strainer cannot be treated as the pressure downstream of it. When pressure is measured directly at the suction flange, upstream losses are already reflected in that reading; do not subtract them twice.
Confirm what the published NPSH curve means
NPSH3 identifies the suction condition at which cavitation causes a three-percent reduction in first-stage total head under the test conditions. It is therefore not a guarantee of cavitation-free operation.
There is also a current terminology detail worth checking. Hydraulic Institute’s explanation of the 2024 ANSI/HI 9.6.1 update says margin assessment now references manufacturer-supplied NPSHR, which must be at least the tested NPSH3. A curve labelled NPSHR should not automatically be assumed identical to NPSH3.
Request the basis of the supplied curve and the required application margin. Avoid applying a universal extra metre or a fixed percentage to every installation. Review the lowest credible NPSHA and the relevant requirement throughout the operating range, including changes in speed and parallel-pump operation.
Work through a positive gauge pressure example
Consider an illustrative chilled-water pump operating with water at 10°C. Assume local atmospheric pressure is measured as 85.0 kPa absolute, suction pressure at the flange is +18.0 kPa gauge, suction velocity is 1.60 m/s and water density is 999.7 kg/m³. The tap is at the NPSH datum, and g = 9.81 m/s².
Use water vapor pressure of approximately 1.23 kPa absolute at 10°C. This value is consistent with the water-property correlation published in the NIST Chemistry WebBook; use appropriate mixture properties if the fluid contains glycol.
Absolute suction pressure = 18.0 + 85.0 = 103.0 kPa. Pressure head above vapor pressure = (103.0 − 1.23) × 1,000 ÷ (999.7 × 9.81) = 10.38 m.
Velocity head = 1.60² ÷ (2 × 9.81) = 0.130 m. NPSHA = 10.38 + 0.130 = 10.51 m.
Apply the stated margin rather than a guessed allowance
For this example only, assume the verified pump submittal gives NPSHR = 8.50 m at the measured duty, and the manufacturer-approved application criterion requires an additional 2.50 m. These numbers are hypothetical, not a manufacturer recommendation for chilled-water pumps generally.
Required NPSHA is therefore 8.50 + 2.50 = 11.00 m. Calculated margin is 10.51 − 8.50 = 2.01 m, leaving a 0.49 m shortfall against the stated criterion. The positive +18 kPa gauge reading does not make this operating point acceptable.
Using 101.3 kPa for atmospheric pressure instead of the measured 85.0 kPa would overstate NPSHA by approximately 1.66 m. A weather service’s sea-level-corrected pressure is unsuitable for this conversion unless corrected back to site pressure. Label the pressure reference in the calculation sheet.
The NPSHR and additional margin are hypothetical. Actual acceptance depends on the selected pump and manufacturer-approved application requirement.
Test the sensitivity before recommending a fix
At the same illustrative flow and temperature, an additional 20 kPa loss upstream of the suction flange would reduce NPSHA by 20,000 ÷ (999.7 × 9.81) = 2.04 m. Available NPSH would become approximately 8.47 m. This sensitivity identifies why a restricted suction path deserves investigation.
It does not predict the actual new operating point. A restriction can change flow, and a variable-speed controller may respond by changing speed. Recalculate using synchronized measurements after any condition changes.
Check warm startup water as well as normal chilled operation. Review the pressure reference connection, suction-path restrictions and instrument accuracy. Increasing fill pressure requires a whole-system pressure-rating review; it should not be used as an improvised cure. Reducing flow also needs an operating-envelope check. Grundfos notes that adequate cooling and lubrication flow must remain when reducing flow to address cavitation.
Use noise as evidence within a broader diagnosis
Air in the system and worn bearings are among the alternative causes of pump noise identified in Grundfos guidance on noisy circulators. A change in sound can direct the investigation, but a recording cannot replace the hydraulic check.
Compare pressure, flow, speed, differential head and vibration at the same timestamp. Note whether the symptom begins during staging, startup or an unusually high-flow condition. If NPSHA is convincingly adequate and the pump is within its permitted operating region, investigate mechanical condition, air and installation effects rather than repeatedly increasing pressure.
A satisfactory calculation also has limits. Poor inlet flow distribution, transient pressure dips and inaccurate instruments can escape a steady-state screen. Persistent symptoms need the manufacturer or a qualified pump specialist, with the measured operating data available.
A field screening checklist
Identify the pump model, impeller, speed and correct NPSHR curve. Record the required margin and its basis.
Verify the suction tap location and instrument reference. Record local absolute atmospheric pressure, temperature and fluid composition.
Measure flow and calculate suction velocity from the actual internal bore. Correct to the manufacturer’s datum when necessary.
Calculate NPSHA with consistent units. Compare the lowest credible result, including measurement uncertainty, with the agreed requirement.
Correlate noise and vibration with the hydraulic record. Escalate a shortfall or unstable operation through the site procedure.
Do not create a cavitation test by restricting the suction valve, deliberately running dry or driving the pump outside its limits. This method supports diagnosis and design review; it is not a field acceptance test or permission to operate through a damaging condition.
Related learning resource
Nexora’s Chilled Water System Design Handbook provides broader context on pump selection, pressure management and commissioning. Pair that design workflow with the selected pump’s certified curves and manufacturer-specific suction requirements.
Sources and further reading
Engineering education only. Verify project conditions, applicable requirements and manufacturer data; use qualified professional judgment before design, selection, operation or maintenance decisions.
Comments