Pump Duty Point Shift After Impeller Trim
A centrifugal pump delivers 45 L/s against 32 m head at 1450 rpm, drawing 22 kW at the shaft. The system curve is approximately H = 12 + 0.0099 Q² (H in m, Q in L/s). The operator trims the impeller diameter from 320 mm to 300 mm to reduce flow, expecting the duty point to move down along the same system curve. After trimming, the measured duty point is 38 L/s at 27 m head, and the absorbed power falls to about 17 kW. The operator is surprised that the flow reduction (16%) is larger than the diameter reduction (6.25%), and that the head fell by more than the affinity-law prediction. Cavitation noise has also appeared at the suction side since the trim.
1. Apply affinity laws to predict the new duty point and compare with the measurement.
Read the worked answer
Affinity-law prediction (diameter change at constant speed):
- Q₂/Q₁ = D₂/D₁ = 300/320 = 0.9375
- H₂/H₁ = (D₂/D₁)² = 0.8789
- P₂/P₁ = (D₂/D₁)³ = 0.8240
Applying to the original duty point (45 L/s, 32 m, 22 kW):
- Q₂ ≈ 42.2 L/s
- H₂ ≈ 28.1 m
- P₂ ≈ 18.1 kW
But the duty point must also lie on the system curve, H = 12 + 0.0099 Q². Solving with the trimmed pump curve is the correct method; the affinity laws only scale the pump curve, they do not fix the intersection.
Check the measured point against the system curve: at Q = 38 L/s, H_system = 12 + 0.0099(38²) = 12 + 14.3 = 26.3 m. The measured head is 27 m, close to the system curve, confirming the duty point moved along the system curve.
Check the predicted point: at Q = 42.2 L/s, H_system = 12 + 0.0099(42.2²) = 12 + 17.6 = 29.6 m, versus the affinity-predicted 28.1 m. The pump curve after trim and the system curve intersect at a lower flow than the naive affinity scaling of the original duty point suggests, because the system curve is not a straight line through the origin - it has a 12 m static component. The measured 38 L/s is consistent with the intersection of the trimmed curve with this system curve.
Conclusion: The 16% flow reduction is not a contradiction. Affinity scaling of the duty point is only valid if the system curve passes through the origin (pure friction). With a 12 m static head, the percentage flow change differs from the percentage diameter change. The correct procedure is to scale the whole pump curve by the affinity laws, then intersect with the system curve.
2. Explain the cavitation symptom and state what must be checked before accepting the trim.
Read the worked answer
Why cavitation appeared. Trimming the impeller reduces the head generated at every flow, but the suction-side geometry and the NPSH required by the pump change as well. Two effects matter:
- The velocity triangles at the impeller inlet change with the reduced diameter and the altered flow, so the NPSH required (NPSHr) generally increases at off-design flow, particularly if the pump now runs further from its best efficiency point.
- The duty point moved to 38 L/s, which may be a lower-flow region where recirculation and higher incidence losses raise NPSHr.
If the available NPSH (NPSHa) was already marginal, the increased NPSHr causes cavitation - the crackling noise, and potentially erosion of the impeller inlet.
Checks before accepting the trim:
- Measure NPSHa at the installation (suction static head, suction line losses, vapour pressure at the water temperature) and compare with the manufacturer's NPSHr curve for the trimmed impeller at 38 L/s. Require NPSHa − NPSHr ≥ 0.5 m margin.
- Verify the duty point is not in a recirculation zone; check the pump curve for the minimum continuous stable flow.
- Confirm the absorbed power (17 kW) against the motor rating and confirm the motor is not now oversized to the point of poor power factor.
- Check shaft vibration and bearing temperature after the trim; a change in axial thrust can follow an impeller trim.
- If NPSHa is inadequate, options are: raise the suction tank level, increase suction pipe diameter to cut losses, or fit a smaller-diameter impeller with a modified inlet, rather than accepting the cavitating condition.