Key takeaway
The best efficiency point belongs to a particular pump curve. It is not a measured efficiency for every operating hour in the archive.
A peak on a chart is not the operating history
A maintenance export lists a pump model, an 82% efficiency figure copied from a brochure and a year of motor-power readings. That combination does not establish 82% pump efficiency throughout the year. The curve may describe a different impeller or speed, and the operating point can move with the system.
DOE’s pump sourcebook identifies head, flow, efficiency, shaft-power requirements, speed and impeller size as information carried by performance curves. ASHRAE describes the operating point as the intersection of pump and system curves. Keep those two ideas together: a correctly identified curve still needs a defensible link to the observed operating conditions.
Build a curve-to-observation map
Consider a hypothetical pump P4. Its retained curve C7 applies to a 200 mm impeller at 1,450 rpm. The chart’s best efficiency point is stipulated as 100 m³/h at 30 m head, with 82% efficiency. A field record instead shows 60 m³/h and 38 m head. No efficiency value at that field point has yet been recovered.
The completed decision retains the field observation and holds its efficiency label. It does not substitute 82%, interpolate from an unseen chart or diagnose wear from the mismatch. The engineer first verifies actual impeller, speed, pressure-tap interpretation, fluid and curve edition. A later curve reading is a model-based estimate with its own source, not a new direct efficiency measurement.
| Evidence | Fictional retained value | What it supports |
|---|---|---|
| Curve identity | C7; 200 mm; 1,450 rpm | One declared pump configuration |
| BEP on C7 | 100 m³/h; 30 m; 82% | Peak at this stipulated curve point |
| Field observation | 60 m³/h; 38 m | An observed flow/head pair in the scenario |
| Efficiency at field point | Not recovered | Hold efficiency assignment |
| Current fitted impeller | Unverified | Hold curve applicability |
Count the time spent away from maximum duty
ASHRAE’s selection discussion considers power at expected operating conditions and the time spent at them. This changes the useful historical artifact from one design-point label to a duty profile. Keep flow, head, speed, control state and power aligned to the same intervals; a peak from one day cannot be paired with average power from another.
A separate hypothetical 12-hour log contains four hours at 20 kW electrical input and eight hours at 10 kW. Energy is 4 × 20 + 8 × 10 = 160 kWh. Assuming the maximum 20 kW for all twelve hours would give 240 kWh, overstating this stipulated log by 80 kWh. The calculation establishes electrical energy under those invented constant-power intervals, not pump efficiency or an achievable saving.
Name the power boundary before comparing
Retain whether power came from an electrical meter, a drive display, a motor estimate or a shaft measurement. DOE’s curve power is expressed as brake horsepower; that boundary differs from electricity entering the drive and motor. A receiver should not compare unlike inputs under one unlabeled power field.
Likewise, head needs the engineering definition actually used. An isolated discharge-gauge reading is not automatically total developed head. Record sensor locations, elevations, relevant velocity treatment and the engineer’s method rather than reconstructing head from an unexplained pressure column. If those records are absent, preserve the raw observation and narrow the proposed analysis.
Return an actionable evidence request
For P4, request the installed configuration and the relevant curve segment. For the 12-hour example, retain the 160 kWh calculation and its two time bins. Do not turn either result into a maintenance-effect claim without a suitable comparison period and operating context.
Inventory can describe curve coverage, interval data and missing configuration. The useful next discussion is whether an evaluator needs duty reconstruction, curve retrieval or a properly designed performance test. The article’s map does not select equipment, certify safe operating limits or establish that a disclosed archive is available for licensing.