The pump needed pressure before it could make pressure.

A pump delivering 60 m³/h at 30 m head was operating with limited suction margin at low tank level. Replacing the pump or changing its foundation would disturb the client’s installation. I compared those options with a simpler intervention: reduce the energy lost before the water reaches the impeller.≈ Assumed

The selected change replaced DN80 suction piping with DN100 and removed high-loss fittings. Calculated NPSH available increased from approximately 3.6 to 5.1 m. The work also included the piping interface, BOM, minimum-level cut-out and low-level commissioning checks.≈ Assumed

2. Separate project facts from the worked example.

The recorded project values establish the design direction. The following illustrative hydraulic model reproduces the rounded 3.6 → 5.1 m comparison; it is not a recovery of the original calculation sheet. Its resistance budgets are chosen explicitly, not inferred as measured fitting data.≈ Assumed

1. Cavitation begins with local pressure.

Water can form vapour when its local absolute pressure falls to the vapour pressure at its temperature. Inside a pump, pressure drops as the liquid accelerates into the impeller eye. Vapour cavities can then collapse as pressure recovers, producing noise, vibration and damage.

NPSH available is the inlet total head above vapour-pressure head. It is a property of the installation. NPSH required comes from the pump’s tested performance at the relevant flow, speed and impeller configuration. These two quantities answer different questions.

For a large, vented supply tank, apply Bernoulli from the free surface to the suction reference plane. Surface velocity is negligible; the pump reference elevation is zero:

Here z is liquid-surface elevation relative to the pump: positive for flooded suction and negative for suction lift. The inlet velocity term is already included through Bernoulli; subtracting it again would double-count it.

7. Turn the pressure budget into a physical layout.

This dimensioned 2,000 mm suction-spool study includes hollow pipe bores, a full-bore valve envelope with handwheel, drilled flanges, fasteners, supports and a flat-top eccentric reducer. Blue identifies piping; green identifies the valve and reducer. It illustrates the layout decisions and does not represent the full installation’s 6 m hydraulic pipe length.≈ Assumed

AB-FS-002 / REV AILLUSTRATIVE ASSEMBLY · mm

Load optional 3D view

Static CAD study. Use the dimension sheet and downloads without JavaScript.

Drag or use arrow keys to rotate. Use the view buttons and zoom slider. Dimensions in mm; illustrative geometry.

The reducer’s smaller centreline is raised by 10 mm to keep the upper internal surface level. A flat-top arrangement can help avoid an air pocket in a horizontal suction run; confirm orientation and straight-run requirements with the pump supplier. A reducer still creates local loss, and an undersized pump nozzle is not made larger by changing the upstream pipe.≈ Assumed

The valve and fasteners are simplified envelopes, not detailed operable internals. The model contains no asserted pressure-rating or flange-standard compliance. Its small outlet section and transitions require their own velocity-based loss treatment in a final calculation; the worked model uses aggregate resistance budgets.

3. The same flow needs less velocity in a larger pipe.

Increasing bore by 25% increases area by 56.25%. The velocity falls to 64% of the baseline. Because velocity head varies with V², its value falls to 40.96%—even before changing a single fitting.✓ Verified by calculation

A nominal DN label is not an internal diameter. Before procurement, repeat this step using the selected pipe schedule and the actual restriction bores. DN100 is a comparison candidate here, not a universal suction-velocity recommendation.✓ Verified by calculation

4. Account for every metre of lost head.

The first term accounts for straight-pipe wall friction; the second accounts for local losses at entrances, valves, bends, strainers and transitions. Use the Darcy friction factor, not the Fanning factor. Each fitting coefficient must use its specified reference velocity.

About 0.565 m of the 1.5 m recovery comes from the straight-pipe term and 0.935 m from the fitting term. Both the larger bore and lower aggregate resistance matter. The K values are illustrative totals; replace them with supplier data or defensible fitting correlations for the actual installation.✓ Verified by calculation

For a detailed check, calculate Reynolds number Re = ρVD/μ, establish roughness ε/D and obtain f from a suitable correlation or the Moody chart. This example holds f at 0.020 to expose the effect of diameter and resistance without pretending the pipe material is known.✓ Verified by calculation

5. Recover inlet margin without changing elevation.

The same atmosphere, liquid temperature and liquid level apply to both cases. The entire gain comes from reducing suction losses. The recovered 1.5 m corresponds to approximately 14.69 kPa of water pressure head.✓ Verified by calculation

The 30 m discharge duty does not belong in this NPSH equation. At the stated duty, hydraulic power is ρgQH = 4.90 kW; shaft power would require pump efficiency. Neither figure establishes the pump’s suction requirement.✓ Verified by calculation

6. Test what happens when conditions change.

Move the controls to inspect the same model at a different flow or tank level. Geometry, vapour pressure and Darcy f remain fixed; losses scale with Q² in this simplified sensitivity study. This is not a pump-curve simulation.

Q = 60/3600 = 0.01666667 m³/s
H₀ = (101325 − 2340)/(998 × 9.81) − 4.510434937 = 5.6000 m
D = 0.08 m: A = 0.00502655 m²; V = 3.315728 m/s; hL = 2.0000 m; NPSHA = 3.6000 m
D = 0.10 m: A = 0.00785398 m²; V = 2.122066 m/s; hL = 0.5000 m; NPSHA = 5.1000 m
Recovery = 5.1000 − 3.6000 = 1.5000 m

A 1 m drop in tank level removes 1 m of NPSH available from both cases. At 72 m³/h, losses increase by 44%: baseline NPSH becomes 2.72 m and revised NPSH becomes 4.88 m at the example level. Raising vapour pressure by 10 kPa would independently remove about 1.02 m from each result.✓ Verified by calculation

Two worked test cases

Case A: Q = 60 m³/h, Δz = 0 m. Case B: Q = 72 m³/h, Δz = 0 m. Both retain all existing inputs. Columns identify the actual hydraulic bore, not a pipe schedule.

QSI = 60/3600 = 0.01666667 m³/s
H₀ = (101325−2340)/(998×9.81)−4.510434937 = 5.600000 m
D=0.08 m: A=π×0.08²/4=0.00502655 m²; V=0.01666667/0.00502655=3.315728 m/s
hL=(0.020×6/0.08+2.0692026781)×3.315728²/(2×9.81)=2.000000 m
NPSHA=5.600000−2.000000=3.600000 m
D=0.1 m: A=π×0.1²/4=0.00785398 m²; V=0.01666667/0.00785398=2.122066 m/s
hL=(0.020×6/0.1+0.9784684314)×2.122066²/(2×9.81)=0.500000 m
NPSHA=5.600000−0.500000=5.100000 m
ΔH = 5.100000−3.600000=1.500000 m
Δp = 998×9.81×1.500000/1000=14.685570 kPa
Ph = 998×9.81×(60/3600)×30/1000=4.895190 kW

QSI = 72/3600 = 0.02000000 m³/s
H₀ = (101325−2340)/(998×9.81)−4.510434937 = 5.600000 m
D=0.08 m: A=π×0.08²/4=0.00502655 m²; V=0.02000000/0.00502655=3.978874 m/s
hL=(0.020×6/0.08+2.0692026781)×3.978874²/(2×9.81)=2.880000 m
NPSHA=5.600000−2.880000=2.720000 m
D=0.1 m: A=π×0.1²/4=0.00785398 m²; V=0.02000000/0.00785398=2.546479 m/s
hL=(0.020×6/0.1+0.9784684314)×2.546479²/(2×9.81)=0.720000 m
NPSHA=5.600000−0.720000=4.880000 m
ΔH = 4.880000−2.720000=2.160000 m
Δp = 998×9.81×2.160000/1000=21.147221 kPa
Ph = 998×9.81×(72/3600)×30/1000=5.874228 kW

Pressure recovery follows Δp = ρgΔH; hydraulic power follows Ph = ρgQH (pressure work per unit time). These are derived from pressure head and continuity. They do not provide shaft power without efficiency. The Q² check follows directly from V = Q/A at fixed geometry, f and K.

Additional exact checks: at 60 m³/h and Δz = −1 m, availability becomes 2.60 / 4.10 m; recovery stays 1.50 m. At 72 m³/h, (72/60)² = 1.44, so losses are 2.88 / 0.72 m. ✓ Verified by calculation

Reynolds-number evaluation and the minimum-level cut-out remain methods to complete, not numerical results: viscosity, manufacturer NPSH requirement, acceptance margin and sensor allowances must be established first.

Pressure/head conversion reference: KSB NPSH. Hydraulic-power reference: KSB pump power relation. The listed formulas and test cases verify arithmetic only.

8. Close the loop at the lowest operating level.

1. Confirm the operating envelope. Obtain minimum tank level, maximum flow, liquid temperature range, site atmospheric pressure and the pump’s NPSH curve.
2. Audit the suction path. Record actual bores, lengths, bends, valve positions, strainer clean/dirty loss and reducer geometry. Check for air ingress, vortexing and trapped-air locations.
3. Set the level cut-out from a margin. Rearranging the equation gives z_min = NPSH_required + margin + h_L − (p_atm−p_v)/(ρg). Add the appropriate allowance for sensor uncertainty and dynamic level change; no numeric setpoint is justified without those inputs.
4. Commission at low level. Log flow, temperature, suction absolute pressure, discharge head, vibration and noise. Convert any gauge suction pressure to absolute pressure before evaluating NPSH.
5. Review and release. Reconcile the calculation, supplier guidance, piping drawing and BOM. Record acceptance criteria and review findings.

Inspect the engineering package.

STEP is an exchange model for CAD import. Downloaded dimensions and numerical examples are illustrative; confirm the actual installation before engineering release.

Technical references

KSB: NPSH · KSB: head-loss equations. References explain the underlying theory; the project comparison and worked-example assumptions are stated above.

Discuss this project with Arpit →

Sources and standards

1. KSB Centrifugal Pump Lexicon: NPSH ↗ — available and required NPSH, installation boundary conditions.
2. KSB Centrifugal Pump Lexicon: Head loss ↗ — continuity, pipe friction and local losses.
3. Existing study calculation data — input provenance and full-precision coefficients.
4. KSB Centrifugal Pump Lexicon: Pump affinity laws ↗ — defines hydraulic output power P = ρgQH; the worked cases hold duty head at 30 m and are not pump-curve predictions.

No governing pump/flange standard or edition has been established by this study. No standards compliance is claimed. The applicable project specification and manufacturer requirements remain to be supplied.

Revision history

3A-R1 · Case-study engine

Reorganized existing content into seven sections. Added input validation, shareable calculation state, two worked checks, accessible equations and a print summary. Engineering inputs unchanged.

Last reviewed

Engineering review not recorded. Populate reviewer, date, scope and evidence only after an actual review.