Measured Leakage Repair Project Records

Measured Leakage Repair Project Records

1. Engineering answer and decision boundary

A defensible measured leakage-repair project record proves which leaks were found, which were actually repaired, how the baseline and post-repair states were measured, and whether the observed change persisted under comparable pressure and operating conditions. It separates direct measurements from engineering estimates and from claims about energy or cost. The record needs an identifiable project, authorised source data, instrument/method evidence, a closed repair register, an as-left test, a persistence check, publication consent and claim approval.

No illustrative arithmetic below is a Reddy Compressors customer result.

2. Six claims that must not be conflated

3. Evidence eligibility, permissions and data integrity

Source chain

Retain the original survey sheets, ultrasonic files where used, pressure/flow/power logs, calibration or verification certificates, work orders, repair photographs and signed acceptance. Log creator, capture date, instrument ID, revision, source owner and immutable evidence reference. A screenshot of a dashboard without raw-period context is secondary evidence.

Identity and confidentiality

Link site and equipment identifiers in a restricted internal record. Public disclosure may use an approved anonymised industry description, but retain internal traceability. Do not reveal production details, security layouts, employee identities or customer trademarks without express authority.

Consent and commercial relationship

Confirm who can authorise use of project results, photos, logo and quotations; document allowed channels, expiry and withdrawal procedure. Disclose relevant service-provider, supplier, dealer or paid relationships. Permission to perform repairs is not permission to publish results.

Evidence maturity

Separate survey-only, repairs completed, short-term measured, adjusted verified and persistence-confirmed records. Never promote a record to a stronger tier simply because the project was invoiced or a photo exists.

4. Measurement boundary and baseline protocol

Declare whether the measured boundary is an individual leak, isolated branch, defined production zone or entire compressed-air station. State whether flow is at actual line conditions or referenced standard conditions; do not interchange m³/min, Nm³/min, SCFM and actual m³/min without declared temperature, pressure, humidity and reference basis. Include the measurement point relative to dryer, receiver, bypass and other consumers.

Record baseline production mix, shift, pressure band, compressor running/loaded/unloaded state, trim controls, receiver dynamics, ambient conditions, instrumentation and known maintenance events. Choose representative baseline windows; exclude or annotate atypical outages, blow-offs and major process changes. A pressure decrease can reduce leakage flow even without repairing a leak, so matched pressure or a supported pressure-normalisation model is essential.

5. Leak survey methods and evidence limits

Ultrasonic identification

Ultrasonic instruments are effective for locating many audible/inaudible turbulent leaks, but displayed leak-flow or monetary values can be algorithmic estimates. Record instrument model, sensitivity, distance, acoustic environment, leak type and estimation assumptions. Do not equate a vendor display to a traceable flow measurement without verification.

Direct capture or branch measurement

For an isolated accessible leak, a suitably rated and safe measurement arrangement may directly quantify flow. Record upstream pressure, temperature, test condition and uncertainty. Do not attach improvised capture devices to hazardous or pressure-containing systems.

Flow logging and isolation

Station or branch flow logging can establish changes when production demand and control states are comparable. Receiver cycling or decay methods need a valid control volume, temperature assumptions, leakage isolation and a safe approved test procedure. A pressure-decay test must not be improvised on a live process.

Power logging

Power meters can verify station electrical input when measurement boundaries and time synchronisation are known. A smaller leak estimate does not imply the same percentage reduction in motor power; compressor unload behaviour, VSD minimum speed and sequencing determine realised savings.

6. Pressure dependence and leakage physics

Many turbulent leaks can be approximated with a compressible-flow model, but geometry, pressure ratio, temperature, discharge coefficient and choking matter. A generic square-root pressure relationship is not valid for every compressed-air leak. For comparisons, prefer measurements at matched pressure or a documented model fitted to the leak class and pressure range. Pressure is absolute in gas-flow equations unless explicitly stated otherwise; gauge pressure cannot be substituted.

A simplified ideal-gas sonic-orifice expression, for an ideal gas under appropriate choked-flow assumptions, is m_dot = C_d A P_up sqrt(gamma/(R T_up)) [2/(gamma+1)]^((gamma+1)/(2(gamma−1))). This is a teaching equation, not an approved field sizing method. Its output depends on discharge coefficient, actual leak geometry, gas composition, upstream absolute pressure and temperature. The critical pressure ratio must be checked before using the choked expression.

7. Repair register and verification chain

8. Normalisation, confounders and causal attribution

Before and after data must be adjusted or qualified for changes in pressure setpoint, product throughput, shift hours, compressor dispatch, seasonal ambient conditions, planned shutdowns, blow-off practices and new consumers. A lower measured flow during lower production is not necessarily leakage repair benefit. Distinguish avoided leak flow under a standardised operating condition from net plant flow change under actual operations.

A project can report a confirmed repair count while holding all savings figures if representative flow or power evidence is missing. Conversely, a credible station kWh reduction can be reported as a measured difference without claiming it was entirely caused by leak repairs where simultaneous controls or production changes occurred. Preserve the unadjusted readings alongside adjusted results.

9. Measurement uncertainty and conservative reporting

Record meter accuracy, repeatability, sampling interval, installation effects, calibration validity, uncertainty due to production matching and model assumptions. For independent differences A−B, a simple standard uncertainty estimate is u_delta = sqrt(u_A² + u_B²); correlated errors require covariance treatment. Do not claim statistical significance or narrow precision without sufficient data. Report intervals, units and method rather than unjustified decimals.

A defensible conclusion may read: “At the stated pressure and test conditions, measured overnight flow fell from A to B within the indicated instrument uncertainty.” It must not silently become “permanent annual electricity savings” without operating-hour, compressor-response and persistence evidence.

10. Persistence and recurrence control

Define the planned reinspection interval according to site criticality, repair population, process vibration, maintenance activity and operating exposure. Reinspect the same tags, sample repaired joints and look for new leaks in adjacent systems. Record the exact post-repair test date and each later check. A single day-one acceptance proves only the observed day-one condition; persistence is a separate evidence claim.

Where repaired leaks recur, retain the original tag and link a new event ID. Analyse failed seals, thread engagement, vibration, hose flexing, thermal cycling and incorrect component specification. Do not overwrite a failed retest with the later successful result. The public claim must reflect the most recent verified status.

11. Worked case A — matched-pressure branch demand (ILLUSTRATIVE)

Fictional teaching dataset: branch flow during matched no-production windows is 2.40 m³/min before repair and 1.55 m³/min after repair, both reported on the same stated reference basis, with pressure held within a documented narrow band. Difference = 0.85 m³/min; relative reduction = 0.85/2.40 ×100 = 35.4%. The 35.4% is a reduction in the measured branch no-production flow for those windows, not an annual plant energy saving. Verify the branch isolation and meter uncertainty before applying the statement to a real case.

12. Worked case B — uncertainty changes the claim (ILLUSTRATIVE)

Fictional values: baseline 2.40 ±0.12 m³/min and after 1.55 ±0.10 m³/min, where ± values are assumed one-standard uncertainties and treated as independent for this teaching case. Difference = 0.85 m³/min; combined u = sqrt(0.12²+0.10²) = 0.156 m³/min. Approximate result: 0.85 ±0.16 m³/min (one-standard uncertainty, not a confidence interval). Real installations require a documented uncertainty budget, including covariance and process variability.

13. Worked case C — repair count is not leakage flow (ILLUSTRATIVE)

Fictional survey: 90 tags identified; 70 repaired; 65 passed immediate retest; 5 remained open and 20 were deferred. Completion rate by tag is 70/90 = 77.8%; immediate verified closure rate is 65/90 = 72.2%. Neither ratio indicates the proportion of leakage flow eliminated because leak sizes differ. The public record may report counts if supported, while holding any flow-savings claim until quantitative evidence is available.

14. Worked case D — annualised air volume is not kWh (ILLUSTRATIVE)

Assume, solely for arithmetic, a verified reference-condition leak-flow reduction of 0.85 m³/min and 4,000 annual operating hours during which the original leaks would otherwise have persisted at comparable pressure. Avoided reference air volume = 0.85×60×4,000 = 204,000 m³/year. This is a scenario volume, not measured annual savings; persistence, demand, pressure and actual runtime must be evidenced. No kWh conversion is justified without compressor-specific incremental power response and controls.

15. Worked case E — incremental power response (ILLUSTRATIVE)

Assume a separately validated system model shows 3.2 kW reduction in average station electrical input attributable to the matched-condition repair intervention, and the site documents 3,500 comparable annual hours. Modelled energy = 3.2×3,500 = 11,200 kWh/year. The result remains modelled unless measured station kWh, baseline adjustment and causal review support verification. If the machine simply unloads for longer with substantial idle power, a direct proportional conversion from flow would overstate benefit.

16. Worked case F — production confounder (ILLUSTRATIVE)

Fictional data show station consumption falling from 700 to 610 kWh per shift, while production fell from 1,000 to 800 units. Specific consumption changes from 0.700 to 0.7625 kWh/unit, an increase rather than an efficiency improvement. This simple comparison does not establish a valid normalised baseline, but demonstrates why raw shift energy is insufficient to attribute a 90 kWh difference to leakage repair. Investigate production, pressure, compressor controls and weather before making a claim.

17. Worked case G — persistence (ILLUSTRATIVE)

Fictional 65 verified closed tags are revisited after an agreed inspection period. Fifty-nine remain closed, four show renewed leakage and two cannot be safely accessed. Observed persistence among inspected tags is 59/63 = 93.7%; among all original closures it is 59/65 = 90.8% confirmed persistent, with two unknown. The denominator and unknown status must be disclosed. Do not describe the entire population as “94% permanent repair success”.

18. Worked case H — cost and emissions boundaries (ILLUSTRATIVE)

Using the purely modelled 11,200 kWh/year from case E and an illustrative energy rate of ₹9/kWh gives ₹100,800/year in energy-charge arithmetic, excluding demand charges, taxes, escalation and tariff structures. With a hypothetical emissions factor of 0.7 kg CO₂e/kWh, modelled avoided emissions would be 7,840 kg CO₂e/year. Neither factor is verified for a site, period or grid. Publish real monetary or carbon claims only with dated approved tariff/emissions factors and clear calculation boundaries.

19. Mandatory asset — Measured Leakage Case Study Template

For each block create repeatable record rows with: field ID; source document ID and revision; measurement or calculation method; value/unit; uncertainty; reviewer; status VERIFIED / PENDING / REJECTED / NOT APPLICABLE WITH REASON; open issue; disposition; approval date. The internal register retains restricted records; the public page displays only approved claims. Empty is not equivalent to not applicable.

20. Example completed record — FICTIONAL / TRAINING ONLY

21. Publication hierarchy and claim examples

Permissible bounded scope

“The authorised project record documents replacement of the specified leaking fittings on the listed branch during the stated period.” Requires actual dated work evidence and consent.

Permissible measured statement

“During matched-condition testing, branch flow changed from A to B at the recorded pressure, with the stated measurement uncertainty.” Requires original matched logs and technical approval.

Conditional modelled statement

“A model estimates an annual reduction of X kWh under Y operating-hour and control assumptions.” Must remain explicitly estimated and must identify source of the model.

Not permissible without evidence

“We eliminated 40% of all factory air losses and guaranteed 25% electricity savings.” A branch flow reduction cannot support a whole-factory or guaranteed energy statement.