Compressor Performance Data Provenance

1. Engineering answer and decision

A compressor performance figure becomes decision-grade only when its source, equipment identity, configuration, measurement definition, reference conditions, test basis, date and applicability are traceable. The same model-family name can appear on an old brochure, an updated datasheet, a quotation and an actual acceptance report with materially different values. It is not a claim that all published data are wrong; it is a safeguard against silently comparing unlike claims.

2. Page ownership and neighbouring boundaries

Cross-link those pages rather than reproducing their whole operating methods.

3. The claim as a controlled data object

Each published duty statement should be a separate claim record, not an unlabeled cell in a comparison spreadsheet. Record claim ID, claimed value, unit, parameter definition, equipment model and suffix, serial or serial applicability range, revision, performance configuration, pressure point, ambient or inlet reference, original source title, document number, revision, issue date, author or issuer, page or figure number, capture date, verification owner, current status and disposition. Link the source file itself. A screenshot without document context may be useful evidence of a claim but is weaker than a complete controlled OEM document.

4. Identity hierarchy and serial applicability

Begin with physical asset tag, manufacturer, full model designation, package serial, airend serial, controller version, motor/engine identity and factory build code. Distinguish a factory specification from an as-installed machine that may have been modified. A serial-range service bulletin may supersede a generic catalogue without changing the marketing model. The applicable range must be checked against the actual serial, manufacture date and revision; a nameplate photograph establishes some identity fields but not the installed airend rotor profile, software settings or the history of an overhaul.

5. What exactly is the performance parameter?

Airflow may mean free air delivery, compressor inlet volume, actual discharge volume, mass flow, normal volume, standard volume, receiver charging capacity or theoretical displacement. Power may mean shaft input, motor electrical input, total package electrical input or station power including dryers and fans. Pressure may mean rated discharge pressure, maximum allowable pressure, test pressure or pressure at a downstream measurement point. Efficiency may be stated as specific power at one point, weighted operating energy or a supplier-defined index. These quantities are not interchangeable without a documented conversion and boundary.

6. Reference conditions: record before converting

For every volume-flow figure, capture the reference pressure, temperature, humidity, gas composition, inlet or discharge measurement location, and whether the supplier uses an applicable named test convention. Record absolute versus gauge pressure explicitly. At altitude, reduced ambient absolute pressure can alter mass intake even if a displayed volumetric flow looks similar. Humidity and water vapour affect dry-gas mass comparisons. Conversion is not a substitute for a missing original definition; if the source omits conditions, label the claim UNRESOLVED rather than selecting convenient defaults.

7. Rated pressure and flow are coupled

The same machine may have different guaranteed capacities at different working pressures. Compare values at the same discharge pressure and with the same regulation/control mode. A brochure heading of “up to 12 bar” does not show that the published maximum flow occurs at 12 bar. An electrical power value at one pressure should not be used to calculate specific power at another. Where only a plotted curve is supplied, digitisation adds reading uncertainty; preserve the original chart and state the interpolation method.

8. Performance source classes

A field logger establishes observed site behaviour only if its sensors, installation and logging period are adequate. Neither a newer date nor a familiar logo automatically makes a source applicable. Rank sources by relevance, authority, traceability and conditions, not by appearance.

9. Source revision and supersession

Maintain the source issue date and revision separately from the date it was downloaded. Compare new editions for changed ratings, motor sizes, pressure points, ambient limits and model suffixes. Record whether the new sheet explicitly supersedes the old one and for which serial range. If a supplier revises a quotation after a configuration change, retain both and mark the earlier quotation as historical rather than deleting it. The decision record must identify which version was actually approved for procurement.

10. Factory data versus site observations

Factory performance values and site readings answer different questions. A factory test may have used clean filters, controlled inlet temperature, calibrated flow measurement and defined loading. A field measurement includes site pressure losses, varying inlet temperature, actual controls, ageing, fouling and measurement error. An observed shortfall should trigger investigation rather than an immediate allegation that the catalogue is false. Conversely, a catalogue value does not prove the currently installed machine is delivering it.

11. Boundaries of package and station energy

When comparing specific power, state whether kW includes the compressor motor only, cooling fan, oil pump, VSD losses, ancillary heaters, dryer, treatment equipment or station-wide auxiliaries. A machine-only number may be useful for a supplier comparison but not for an energy-savings forecast that includes the whole station. Avoid combining electrical power from one boundary with delivered flow measured across another. Record synchronous power and flow samples where possible and the operating state to which they relate.

13. Time-series and control-state evidence

Performance claims from variable-speed machines should identify speed, load fraction and pressure. A single average can conceal long unloaded periods. Segment operating records into loaded, unloaded, stopped, standby, transient and fault states before deriving typical power. Note missing samples, sensor saturation, time-zone offsets and clock drift. Where a supplier uses a duty-cycle weighted value, obtain the weighting distribution. A reported annual efficiency gain without a baseline demand profile is a scenario assumption, not a verified site outcome.

14. Duty applicability and derating

Record ambient range, altitude, cooling-air or cooling-water design, gas composition, filtration state, inlet restrictions, electrical frequency and site duty cycle. A rating based on 50 Hz must not automatically be applied to a different speed/frequency configuration. Portable compressors may have engine power and derating constraints that differ from stationary electric packages. Oil-free and specialist gas systems require separate material, purity and safety approvals. Any proposed extrapolation beyond supplier data must be reviewed by the competent designer and labelled as such.

15. Reconstruct the original claim lineage

For a disputed number, follow the chain: website or sales slide → brochure or datasheet → exact OEM document revision → test report or calculation basis → serial/configuration applicability → signed project submittal → as-built record. Capture every transformation, such as unit conversion, rounding, interpolation or replacement of a value after a design revision. If a step cannot be reproduced, the derived number remains a traceable assertion but not a verified technical result.

16. Data conflicts and resolution protocol

Conflicting values should not be averaged. First determine whether they refer to different pressures, units, machine variants, revisions, test standards or scope boundaries. If the discrepancy persists, issue a technical query identifying both source IDs, the exact conflicting values and the decision at risk. Ask the issuing authority for a signed clarification or revised data sheet. Preserve the superseded record, the answer, the responsible reviewer and the final accepted value with date. A procurement deadline is not evidence of technical equivalence.

17. Uncertainty and significant figures

Every measurement and conversion carries uncertainty. Avoid reporting a calculated value to five decimals when the input chart was read approximately. If comparing two close supplier ratings, quantify whether the difference is material relative to measurement uncertainty and operating variability. A 2% apparent advantage cannot be treated as decisive if both datasets use materially different boundaries or an uncertainty of similar magnitude. Where uncertainty is unknown, record it as an evidence gap rather than inventing a tolerance.

18. Performance maps and interpolation

For a compressor with multiple speed or pressure points, retain the complete map and its axes, not a single extracted number. Interpolation between validated points may be appropriate if the supplier permits it and the method is stated; extrapolation beyond the tested envelope is not equivalent. For centrifugal machines, surge, choke and gas composition boundaries may govern; for reciprocating machines, speed, clearance and temperature matter; for rotary screw machines, pressure, speed and control strategy matter.

19. Acceptance versus marketing language

Terms such as “up to”, “nominal”, “typical”, “guaranteed”, “measured”, “tested” and “estimated” should be retained verbatim in the provenance record. A nominal performance curve may guide preliminary screening but should not become a contractual guarantee by copying it into a tender table. A guaranteed point is only as strong as its specified test conditions, tolerances, exclusions and acceptance protocol. The project technical authority decides which statements can be relied on for selection and which require verification.

20. Technical hold points and decision statuses

Use statuses CAPTURED, UNVERIFIED, APPLICABILITY CHECK, CONFLICT, VERIFIED FOR STATED BASIS, SUPERSEDED, REJECTED and APPROVED FOR PROJECT USE. “Verified for stated basis” is narrower than “approved for this site.” Require explicit owner approval for project use, and block downstream selection where a critical value remains disputed. Do not allow a website form or automated matching rule to promote a claim to approved status without human evidence review.

21. Service history and as-found condition

Performance provenance should preserve evidence of airend exchanges, cooler replacement, motor rewinds, firmware changes, setpoint adjustments, filter restrictions and repairs. The as-built or as-left configuration may differ from the original OEM build. Attach maintenance work orders and commissioning reports with dates and signatures. Such records can explain a change in observed performance but do not automatically revise the original factory rating. Keep historical and current equipment states separately identifiable.

22. Supplier enquiry and clarification package

Send the supplier a compact query with complete model/suffix, serial, revision, duty, claimed figure, source page, disputed definition, proposed site conditions and the exact decision being made. Ask for a controlled response with the issuing engineer, applicable model/serial range, guaranteed or indicative status, reference conditions, document revision and any derating. Where suppliers decline to disclose proprietary test data, record the limitation and seek an acceptable witnessed or contractual verification path.

23. Procurement and warranty boundary

The purchase order should cite the accepted performance document and revision, defined guaranteed points, applicable test protocol, deviations, accessories included, acceptance responsibility and remedies. A warranty statement is not a substitute for performance evidence. Where a legacy machine or third-party replacement is involved, do not infer compatibility or continuing warranty from an OEM family name. Retain signed change notices and final as-built supplier submittals.

24. Commissioning and field verification

Before test, freeze the instrumentation plan, operating states, calibrated equipment, test duration, gas conditions, boundary points, control settings and permissible tolerances. Record as-found and as-left conditions and any corrective work. Compare field results to the appropriate guarantee only after adjusting or qualifying the reference basis by an approved method. If acceptance cannot be established because the test point was not achieved, report “not demonstrated” rather than a pass or fail inferred from unrelated readings.

25. Data governance and audit trail

Assign a unique claim ID and immutable attachment hash or controlled file identifier. Preserve original units, exact quoted wording, source file, OCR/extraction method if used, editor, review date and all derived fields. Separate the source value from converted display values. An administrator may correct a transcription error but must not silently overwrite historical evidence. Export the full register, revision history and unresolved issues for the project dossier. Apply access controls to confidential OEM and customer files.

26. Visual website implementation

The public article should display a clear explanation of evidence tiers, sample register and worked comparisons. Use responsive tables with column headings and mobile-friendly overflow; show units beside each value. A demonstration lookup must be labelled as an illustrative tool, not as an OEM-backed performance database. Do not display customer serial numbers or proprietary test documents without permission. Provide an accessible downloadable blank register, if the developer elects to implement it, with input/evidence, method, status, owner, open item and disposition fields.

27. Quantitative engineering examples — illustrative, not OEM claims

Case A — equal catalogue flow, unequal reference basis

Supplier A lists 600 m³/h FAD at 7 bar(g) using stated inlet reference conditions. Supplier B lists 600 Nm³/h with an unspecified normal basis and a different pressure point. The two numbers are not directly comparable. Register both exactly as published, record the missing normal reference and pressure basis for B, and issue a supplier query. Disposition: HOLD comparison until common reference and guaranteed pressure point are documented.

Case B — apparent specific-power advantage

At the same pressure and flow reference basis, candidate X is quoted at 75 kW package input and 12.5 m³/min delivered flow: 75 / 12.5 = 6.00 kW per (m³/min). Candidate Y lists 72 kW motor-shaft power and 12.0 m³/min: 72 / 12.0 = 6.00 in arithmetic terms, but the power boundaries differ. Disposition: no energy ranking until both power values use the same package boundary and operating state.

Case C — revision conflict for one family

A 2020 family brochure reports 11.2 m³/min at 8 bar(g). A 2025 sheet reports 10.6 m³/min at the same pressure for a revised suffix. The original machine carries an older serial and a documented replacement airend. Neither value can be adopted from model-family identity alone. Obtain serial-specific factory build and exchange records, then record which curve is applicable. Disposition: configuration-specific verification required.

Case D — pressure units and absolute reference

A record states 8 bar but omits whether gauge or absolute. At atmospheric pressure approximately 1.013 bar absolute, 8 bar(g) is approximately 9.013 bar absolute; 8 bar absolute is approximately 6.987 bar(g). The difference is material to duty and pressure ratio. Disposition: reject unqualified conversion; seek original author clarification and record site atmospheric basis.

Case E — chart interpolation

A supplier curve reports 12.0 m³/min at 7 bar(g) and 11.4 m³/min at 9 bar(g). Straight-line interpolation at 8 bar(g) would yield 11.7 m³/min, but only as a calculation between two points. Disposition: provisional screening only until supplier confirms that interpolation is valid for the selected unit.

Case F — field test versus factory curve

A site test reports 10.1 m³/min while a factory sheet gives 10.8 m³/min. The field test was conducted with a fouled intake filter, higher inlet temperature and different measurement location. A 6.5% numerical difference alone does not prove degradation. Document instruments, actual pressure, operating state, inlet conditions, correction authority and maintenance findings. Disposition: investigate and retest on a controlled basis.

Case G — energy extrapolation

An audit report assumes a 5 kW saving for 6,000 operating hours, producing 30,000 kWh/year. That arithmetic is correct only if the 5 kW difference persists for all 6,000 hours. A duty profile showing 3,000 loaded hours and 3,000 unloaded hours requires state-specific power differences. Disposition: label the simple annualisation as a scenario, not verified savings; obtain time-weighted data.

Case H — test instrument scaling

A current-transformer ratio of 400:5 corresponds to a factor of 80. A logger configured with 300:5 applies factor 60, underreporting measured current by 25% relative to the correct scaling, all else equal. Record the installed CT marking, meter settings and calibration record. Disposition: invalidate affected power trend pending correction and repeat or justified reconstruction.

Case I — same family, different frequency

Two compressors carry the same family label; one is a 50 Hz fixed-speed build and the other a 60 Hz build with a different motor and pulley arrangement. Airflow and input power cannot be copied between them from the shared family name. Disposition: retain two separate configuration records and require the exact OEM curve for each.

Case J — source date versus download date

A datasheet downloaded in October 2026 carries an internal issue date of 2018 and no revision number. A 2024 service bulletin may have changed permissible operating settings for some serials. The download date does not make the sheet current. Disposition: mark historical pending OEM revision check and attach the supersession inquiry.

29. Worked controlled comparison: two plausible records, one family

Review outcome: both records are CAPTURED but project-use approval remains HOLD. Record A may proceed to an applicable guarantee and site duty review. Record B requires the original controlled OEM sheet, replacement airend record and revised performance basis. Do not combine the highest flow from A with the higher pressure from B as if it were a single machine rating.

30. Release workflow and acceptance tests

Capture

Register exact claim wording, source attachment, source date and units; assign a claim ID.

Match identity

Verify complete model, serial, revision, as-built configuration and modifications.

Normalise carefully

Only convert quantities after original reference definitions are established; retain originals.

Resolve conflicts

Log discrepancies and seek signed supplier or responsible-engineer clarification.

Verify project duty

Check pressure, flow, quality, ambient, controls and power boundary against actual duty.

Accept or hold

Issue a named disposition with evidence, restrictions, test plan and approval signature.

Negative tests for a digital register: missing serial; family-name match only; gauge/absolute ambiguity; FAD-versus-normal-volume ambiguity; mixed shaft and package power; absent source revision; conflicting datasheets; superseded source; expired calibration; unknown controller revision; missing attachment; invalid formula; unapproved extrapolation; open critical query; unauthorised approval; export that drops provenance. Each must fail closed or remain visibly UNVERIFIED.

31. Engineering FAQs

Can two brochures for the same model both be correct?

Yes, if they apply to different revisions, pressures, reference conditions or package configurations.

Is the latest PDF always authoritative?

No. Check issue date, revision, serial applicability and supersession.

Does a nameplate show delivered airflow?

Often it does not; identify what it actually marks and consult applicable performance data.

Can kW per m³/min be compared across suppliers?

Only when flow basis, pressure, electrical boundary and operating state align.

What if the OEM will not provide a test report?

Record the limitation and seek an agreed guarantee or witnessed acceptance basis.

Does field data override a factory guarantee?

Not automatically. First reconcile measurement conditions, equipment state and test method.

Can a marketing “up to” value be used for design?

Not as a guaranteed point without supporting conditions and approval.

How is a source conflict resolved?

Identify exact documents, values and differences; request a controlled technical clarification.

What if the serial number is unreadable?

Hold serial-specific claims until alternate OEM-accepted identity evidence is established.

Is a downloaded file date the document issue date?

No. Record both separately.

May missing reference conditions be assumed?

No. Mark the claim unresolved and obtain the issuer’s definition.

What does verified for stated basis mean?

The claim is supported under its documented conditions, not necessarily approved for the site.

How should chart values be recorded?

Retain the chart, reading method, interpolation restrictions and uncertainty.

What is the mandatory asset?

The Performance Data Provenance Register with evidence, method, status, owner and disposition.

Does a service replacement affect published performance applicability?

It can; document the installed revision and seek an applicable performance basis.

What belongs in final handover?

Approved claims, source copies, revisions, calculations, tests, open items and as-left configuration.

Can the website show live guaranteed specifications?

Only with maintained authoritative data and explicit approvals; otherwise show guidance and templates.

Can apparent energy savings be presented as proven?

Only when supported by appropriate measured baselines, operating profiles and verification.