Power meter data collection should prove whether compressor input kW and kWh actually fell after compressed-air leak repairs—not merely add another dashboard. A Thai factory can find one hundred leaks yet deliver little electrical reduction if compressors continue to run unloaded. Investment-grade evidence therefore links the repair timestamp, machine state, flow, pressure and comparable production to the same electrical clock, while keeping estimated opportunity separate from verified saving.
The primary search intent of this guide is how to design power meter data collection for repair verification, a 30-day PoC, RFP, FAT/SAT, acceptance and ROI within a compressed-air system. Ultrasonic survey and repair are treated as prerequisites for the electrical evidence, not as a general leak-detection tutorial. DOE’s 20–30% statement remains general context, not a fact or guarantee for the reader’s site. Safety comes first: follow factory EHS, risk assessment, isolation/LOTO, depressurisation, competent-person and OEM requirements, and never touch, tighten or dismantle a live pressurised component.
Use power meter data collection to close the repair-verification loop
A typical campaign surveys the plant during a shutdown, attaches tags, exports a list and hands it to maintenance. Management, however, needs more than a tag count. It needs the annual addressable cost, the approved repair window, proof that pressure quality remained acceptable, evidence that each repair passed a re-test, and confirmation that the investment produced an attributable result.
ISO 11011:2013 frames a compressed-air assessment from energy input through the useful work delivered. It divides the system into supply, transmission and demand and covers assessment roles, analysis, reporting, documentation and estimated savings. This is the right boundary for a leakage programme: do not inspect couplings in isolation from compressor controls or end-use demand.
| Subsystem | Typical scope | Key leakage question |
|---|---|---|
| Supply | Compressors, dryers, receivers, sequencing | Did repairs reduce loaded time, input kW or the number of machines running? |
| Transmission | Headers, branches, valves, hoses, fittings, drains | Where are material leakage and pressure losses located? |
| Demand | Production equipment, blowing, instruments, fixtures | Can leakage be separated from legitimate continuous use? |
The closed loop is: pre-repair power meter data collection → register → approve and repair → re-test → post-repair collection → verify flow, pressure and power → update the standard. A tag without an owner, a verbal completion claim, or a spreadsheet that adds estimated leak costs without measured power verification is not investment-grade evidence.
Why the DOE 20–30% figure is not your savings target
DOE’s 2004 Compressed Air Tip Sheet #3 says that leaks can waste 20–30% of compressor output and describes 5–10% of total system flow as a typical cost-effective leakage target. Those statements are useful for screening, but the age and broad scope of the document matter. A Thai plant may have different pressures, controls, maintenance, end uses and operating hours.
Use 20–30% to explain why measurement deserves attention—not to promise that the plant will save 30%. Set a PoC target from non-production flow, load/unload cycles, measured input power, header pressure, survey evidence and repair history. The sourcebook advice to test at least quarterly and treat leakage above roughly 10% as an improvement opportunity is also a technical benchmark, not law and not a universal acceptance threshold. The site’s recurrence rate and operational risk should set the long-term frequency.
Combine power meter data collection with ultrasonic, flow and pressure evidence
The purpose of air-leak detection is not to find a sound. It is to convert a location and severity into a safe repair order, then verify the response of the compressed-air system. No single measurement gives location, flow, electrical impact and operational consequence at once.
| Method | Best use | Limitation | Role in a PoC |
|---|---|---|---|
| Ultrasonic detector | Locating high-frequency leak signatures in noisy areas | Reading depends on device, range, distance, pressure and surroundings | Locate, tag and re-test |
| Off-production flow | Quantifying residual system or zone demand | Legitimate continuous uses must be separated | Baseline and zone priority |
| Pressure trend | Identifying instability and remote-point shortfall | Pressure alone does not prove leak flow or energy | Quality KPI and side-effect check |
| Power-meter data | Verifying input kW/kWh after repair and control changes | Needs compressor state and production context | ROI and acceptance |
| Load/unload or VSD status | Explaining control response | Signal definitions and OEM logic must be checked | Diagnose why a repair did not reduce power |
Ultrasonic instruments are effective for narrowing down a location. DOE also describes high-frequency acoustic detection as a practical method. Yet a dB value is not automatically a universal flow rate. An RFP should require the instrument model, inspection or calibration status, distance, angle, system pressure, sensitivity, conversion method and uncertainty class. Important leaks should be cross-checked through zone flow or system response where feasible.
Turn every leak tag into a maintenance work order
A usable leak record should include:
- unique Leak ID and QR code;
- building, line, asset, piping system, photograph and drawing location;
- discovery time, inspector, instrument, setting, distance and pressure;
- estimated flow, calculation method and confidence category;
- provisional annual cost with all assumptions visible;
- safety/access class, isolation point and required shutdown window;
- production, quality and safety criticality;
- repair owner, due date, materials and work-permit reference;
- before/after test result, photo, replaced part and approver;
- post-repair effect on flow, pressure and compressor power.
A tag with no exact location, owner or shutdown path will remain open. Define the next action and closure criteria at creation. If CMMS integration is not yet possible, the leakage register still needs an accountable owner, due date, re-test result and a common ID between the map, repair order and dashboard.
Prioritise annual loss and criticality—not loudness
The loudest apparent leak is not necessarily the highest-cost job. Device readings change with distance and settings, and a leak on a header pressurised 24 hours a day has a different annual cost from one on a fixture used for two hours. A smaller leak that destabilises pressure at a quality-critical process may deserve higher priority than a larger, isolated leak.
A practical priority model separates financial and operational factors:
Priority = annual loss class + production impact + quality impact + safety/access + repairability + recurrence
The factory must set the scoring weights. Safety cannot be traded against ROI, so serious EHS conditions need a separate gate. Damaged hoses or unstable components should trigger site EHS escalation, area control and an approved isolation decision—not an attempt to complete the survey.
DOE’s educational example shows a small number of larger equivalent holes accounting for most of the calculated saving. It is not a site distribution, but it illustrates why large leaks should normally be repaired first. A Thai plant can use categories such as A for safety/production critical, B for high annual loss with an available repair window, C for planned-maintenance work, and D for further measurement.
| Priority | Example decision | Target action |
|---|---|---|
| A | EHS concern, damaged hose or critical pressure risk | EHS control and competent repair after approved isolation |
| B | High annual loss and repairable in a short shutdown | Repair and re-test at the earliest approved window |
| C | Moderate cost, scaffold or equipment shutdown needed | Integrate with the next PM/planned stop |
| D | Small or uncertain, possibly legitimate demand | Measure again and review during the next route |
Power meter data collection design: align the compressor-state clock
“We fixed the leaks but power did not fall” is a common outcome. When demand drops, a load/unload compressor may continue consuming substantial unloaded power. VSD machines and multi-compressor sequencers also respond differently. Therefore, power-meter data collection must include input kW and kWh together with run, load, unload, VSD speed or load percentage, discharge pressure, header pressure, flow, dryer state and relevant production context.
A useful data hierarchy is:
- utility bill or incoming supply data for financial reconciliation;
- compressor-room feeder power for the system input;
- individual compressor power and state for sequencing analysis;
- header flow, pressure and dew point for delivered service;
- zone flow or non-production flow for leakage segmentation;
- output, shift, product and downtime for baseline adjustment.
Clock misalignment can reverse cause and effect. If a power meter, PLC and flow meter record the same change several minutes apart, the system may appear to reduce power before a machine unloaded. The RFP should specify time synchronisation, Asia/Bangkok timezone, missing-data treatment, quality flags, resend logic, engineering units and totaliser resets. Use faster raw data for control analysis and lower-frequency aggregates for long-term reporting.
For general leak-finding, prioritisation and recurrence prevention, use the existing foundation article Compressed-Air Leak Detection and Repair. This article focuses on the next stage: accepting repair performance through electrical evidence. For wider TCO planning, see Factory IoT Cost 2026; for zone utility measurement, see Flow Meter Monitoring: Thailand PoC, RFP and Acceptance.
A 30-day power meter data collection PoC for before/after proof
A 30-day PoC is not a compressed-air transformation for the whole site. It should prove one closed loop for a compressor room and one or two zones: baseline, survey, safe repair, re-test and measured power response. If the 30 days do not include comparable production, extend the evidence window or hold acceptance. Do not force a comparison between dissimilar operating conditions merely to meet the calendar.
Days 1–5: freeze scope, safety and baseline
- Confirm compressors, headers, zones and end uses on a P&ID or simplified system drawing.
- Approve a RACI covering EHS, maintenance, production, electrical, energy and IT/OT.
- Define what non-contact observations are permitted while running and the LOTO/depressurisation process for repair.
- Review 12 months of bills, the electricity contract and the operating calendar.
- Check power, state, flow, pressure and production-context data quality.
- Agree on the baseline period, exclusions, adjustments and acceptance KPIs.
Do not promise a leakage or saving percentage at this point. First define what constitutes comparable conditions. Non-production flow is useful only after legitimate instrument air, purge, safety and 24-hour demand have been identified.
Days 6–12: perform the ultrasonic route and zone balance
A competent surveyor follows an approved route and makes non-contact ultrasonic observations. Each finding receives a Leak ID with location, pressure, range, settings, photograph, estimation method and safety class. Where operationally and safely feasible, zone flow or approved isolation tests help quantify residual demand. If the survey indicates a potentially unsafe condition, stop and use the site’s EHS escalation process.
Days 13–20: approve and execute the repair batch
Review estimated annual cost together with materials, skills, scaffold, cleanliness, downtime and quality risk. Repairs are performed only after the approved work permit, LOTO, isolation, depressurisation and zero-energy confirmation, by competent personnel following OEM and factory standards. Do not tighten a live joint. Re-pressurisation and functional checks must also follow the approved procedure.
Days 21–26: re-test and review supply control
Re-test repaired points under comparable conditions. If flow falls but the same compressors remain online at the same pressure, review sequencing or control with the OEM or a qualified specialist. DOE’s tip sheet also advises re-evaluating supply after leaks are repaired so run time can match reduced demand. Any setpoint or logic change, however, must consider pressure at the furthest user, dew point, starts per hour, motor loading, warranty and change control.
Days 27–30: decide acceptance and scale-up
Acceptance reviews data completeness, repair re-tests, pressure stability, adjusted kW/kWh and unresolved risk—not only the number of tags closed. If a measurable benefit is not confirmed, classify the reason: inadequate repair, control non-response, non-comparable conditions, measurement error or legitimate demand misclassified as leakage. A transparent “no verified saving yet” result still improves the design of a wider programme.
Write a supplier-comparable power meter data collection RFP
“One ultrasonic camera” and “one dashboard” do not define comparable scope. Suppliers will estimate leakage differently unless the RFP fixes the evidence, workflow and acceptance method.
| Requirement area | RFP content | Required evidence |
|---|---|---|
| Scope | Compressors, zones, pipe length, access limits, hours | Asset list, diagram and exclusions |
| Safety | Training, permit, non-contact route, LOTO, depressurisation, competence | Method statement, JSA, qualifications, approvals |
| Ultrasonic | Device, calibration/check, distance, setting, pressure, conversion | Raw data, photos and Leak ID register |
| Measurement | Power, state, flow, pressure, dew point, production, period, accuracy | Instrument list, certificates, tag list, clock test |
| Data | Retention, missing data, resend, quality, CSV/API, access | Data dictionary, sample export, backup test |
| Workflow | Discover, approve, repair, re-test, close and recur | RACI, statuses and audit trail |
| Verification | Baseline, adjustment, exclusion, control response, tariff | M&V plan, calculation workbook, assumption ledger |
| Handover | Thai/English training, SOP, spares and support | Training record, manuals and SLA |
Compare TCO, not only initial price. Include recurring survey, calibration, communication, software, storage, support, repair materials, shutdown labour, scaffold and training. If the instrument automatically calculates flow, require an auditable conversion basis and exportable raw readings. Leak IDs and repair history should remain portable if the supplier changes.
OT cybersecurity and change control
Connecting meters and gateways introduces OT requirements even in a leakage project. Specify read-only acquisition, segmentation, least privilege, certificate/key management, time sync, audit logs, vulnerability handling and controlled remote support. Any write access to compressor controls should be separated from monitoring and require OEM approval, change request, backup, rollback and FAT/SAT. A short PoC is not justification for an unmanaged personal laptop or mobile router to become permanent infrastructure.
FAT/SAT should test continuity of the electrical evidence
FAT should inject simulated tags and test calculations, alarms, workflow, reports and access rights before site installation. Include missing values, outliers, clock drift, totaliser reset and duplicate events. A Leak ID must link its before/after evidence, and an unapproved repair should not be closable without the correct role.
SAT uses real instruments and equipment states. Check phase/CT ratio, units, time, totaliser direction and feeder reconciliation. Confirm that run/load/unload signals match the physical compressor, that flow and pressure ranges are plausible, and that stale values are not displayed as healthy after communication loss. Trace a Leak ID across the physical tag, photograph, drawing, CMMS order and dashboard.
| Test | Passing evidence | Failure example |
|---|---|---|
| Completeness | Acquisition meets the agreed threshold and gaps are classified | Gaps are filled with zero to create a false reduction |
| Time sync | Power, state, flow and pressure align within tolerance | kW appears to change before the control event |
| Meter reconciliation | Feeder and machine meters agree within the agreed band | Wrong CT ratio or reversed totaliser |
| Leak ID trace | Discovery through re-test is uniquely linked | Photo and repair ticket refer to different locations |
| Safety evidence | Permit, LOTO, depressurisation and competence are complete | Scope assumes repair under pressure |
| Verification | Adjusted measured kW/kWh are compared on like conditions | Estimated tag value is booked as actual saving |
Power meter data collection acceptance KPIs: separate estimate from outcome
Do not put every metric into one “leak saving” number. Keep leading, execution and outcome KPIs distinct.
Leading KPIs
- route/zone coverage;
- data completeness and clock synchronisation;
- mandatory fields completed for each Leak ID;
- percentage of estimates with pressure, range and method recorded.
Execution KPIs
- on-time repair rate for A/B tags;
- percentage of repairs re-tested;
- pass rate at re-test;
- recurrence and mean repair lead time;
- completeness of permits and LOTO records.
Outcome KPIs
- change in non-production zone flow;
- adjusted compressor kW/kWh by comparable production;
- header and remote-point pressure stability;
- loaded/unloaded time, machines running and VSD load;
- avoided cost recalculated with the actual bill structure.
ISO 50001:2018 publicly describes a PDCA management framework using energy performance indicators (EnPIs) and energy baselines (EnBs), and it is technology-neutral. Leak registers, power trends and verified repairs can support that framework, but installation of meters or detectors does not confer ISO 50001 certification.
DEDE’s 2025 sample Energy Management Manual includes an air-compressor example with equipment configuration, intended use, discharge-pressure management and controlled preparation/review/approval fields. This supports the principle that setpoints and responsibilities should be documented. It is an example, not a mandatory template; never copy its pressure or equipment values into another factory without engineering review.
Use Thai electricity tariffs carefully in ROI
PEA’s structure effective for September 2026 billing lists, for Type 3.2 below 22 kV, a demand charge of THB 210.00/kW, peak energy at THB 4.3297/kWh and off-peak energy at THB 2.6369/kWh. These figures apply only to that contract category and voltage, and Ft and VAT are separate. ERC lists retail Ft for September–December 2026 at 16.23 satang/kWh. It is therefore wrong to tell every plant that its electricity price is THB 4.3297/kWh or to add Ft twice.
Start with the actual bill: contract category, voltage, TOU, demand, Ft, VAT, power-factor items and minimum-charge conditions. A kWh reduction has different value by time period. It may not lower demand charges unless it coincides with the billed maximum. Capacity released or a compressor purchase deferred should be shown as a separate benefit from energy and demand savings.
Illustrative ROI model
The following is a TOMAS TECH teaching model, not a quote, benchmark, site result or guarantee.
| Item | Assumption | Calculation |
|---|---|---|
| Representative compressor input | 180 kW average | Measure with meter and state signals |
| Operating time | 6,000 h/year | Replace with the actual calendar |
| Attributable reduction after repair/control | 12 kW | Requires an adjusted measured baseline |
| Illustrative blended energy value | THB 3.80/kWh | Rebuild from actual bills |
| Annual gross benefit | THB 273,600 | 12 × 6,000 × 3.80 |
| Initial programme cost | THB 420,000 | Assumed PoC, measurement, survey and repair |
| Annual recurring cost | THB 72,000 | Assumed survey, calibration, data and support |
| Annual net benefit | THB 201,600 | 273,600 − 72,000 |
| Simple payback | 25.0 months | 420,000 ÷ 201,600 × 12 |
The critical 12 kW is not the sum of label estimates. It is the measured input reduction attributable under comparable conditions. A load/unload machine may simply spend longer unloaded unless sequencing changes. A VSD plant can also be constrained by minimum speed, machine interaction or pressure settings. Use sensitivity cases for kW, hours and tariff, then let finance add tax, funding, depreciation and shutdown cost. Keep NPV, capacity headroom, pressure quality and maintenance burden as separate decision dimensions.
Prevent recurrence after the first repair wave
Leaks recur through vibration, temperature, work changes and ageing. Move the programme from an annual event into standard work:
- safe daily visual/audible observation without contact;
- monthly review of open Leak IDs, high-priority work, pressure, off-production flow and energy intensity;
- periodic ultrasonic routes, using the DOE quarterly suggestion only as a starting reference;
- change control for new fixtures, hoses, blowing uses and drains;
- planned-stop batch repair for items requiring scaffold or isolation;
- annual review of EnPI/EnB, tariffs, capacity, calibration, training and supplier performance.
Procurement standards also matter. Rationalise approved hoses, couplings, fittings and drains; define installation, support, vibration control and replacement standards. Do not respond to every remote pressure problem by raising compressor discharge pressure. Measure distribution loss, filter differential pressure and genuine peak demand first.
Provide Thai and English SOPs so that discovery, shutdown request, repair, re-test and closure survive shift changes. Management reports should display discovered estimate, repaired estimate, verified electrical benefit and unresolved risk separately. Never add estimated and verified savings in the same total.
Frequently asked questions
Will buying an ultrasonic detector reduce energy cost?
The detector helps locate problems. Savings require repair authority, shutdown access, materials, re-test, control response and power verification. Test data export, conversion basis, calibration, training and register integration during the PoC before choosing a tool.
Can an ultrasonic dB reading be converted directly to leak flow?
Only with a defined method and its assumptions for instrument, range, sensitivity, frequency, pressure and environment. Record those factors and cross-check material cases with zone flow or system response.
Is all non-production flow leakage?
No. Instrument air, purge, safety uses, continuous machines and drains may be legitimate. Inventory them before isolating zones. Stopping a legitimate use can create a quality or safety event.
What sampling interval is needed for power meter data collection?
It depends on the decision. Sequencing and load/unload analysis need faster raw data; a monthly ROI view may use 15-minute aggregates. Accuracy, CT ratio, time sync and gap treatment are more important than high speed alone.
Should a plant set 20–30% as the savings target?
Not for budget approval. It is an old general DOE observation. Use measured leakage, power response, operating hours and the actual tariff to set the site target.
Is the PoC a failure if power does not fall after repair?
Not necessarily. Pressure stability or capacity may improve while compressor control fails to respond. But do not book electrical savings until measured kW/kWh falls under comparable conditions. Treat control optimisation as a controlled next step.
What are the most important RFP acceptance conditions?
Trace each finding through repair and re-test with one Leak ID; maintain complete safety evidence; align power, state, flow and pressure clocks; and verify outcomes under comparable production. A polished screen and a high tag count are not enough.
Can inspection or repair be done while the system runs?
An approved non-contact ultrasonic route may be conducted during operation under the site’s risk assessment and access rules. Stop if a hazard is suspected. Touching, tightening, opening, applying liquids or climbing is not automatically safe. Repairs should be done by competent personnel after approved isolation, LOTO, depressurisation and zero-energy verification, following site and OEM procedures.
Conclusion: accept leak repairs through power meter data collection
Investment-grade power meter data collection connects a repairable Leak ID and safe shutdown to before/after re-tests, synchronised compressor state, flow, pressure, measured kW/kWh and transparent ROI. Close one evidence loop in the 30-day PoC; do not book DOE’s 20–30% context or tag estimates as actual savings. Apply PEA tariffs only after confirming the contract and bill. General leak detection remains in the foundation article, while this guide centres on comparable data quality and acceptance. Safety precedes ROI at every stage.
TOMAS TECH can help define power meter data collection points, time synchronisation, the PoC boundary, acceptance KPIs, meter architecture and a supplier-comparable RFP around the compressors already installed at your Thai factory. Contact TOMAS TECH
References
- ISO 11011:2013 — Compressed-air energy-efficiency assessment
- US DOE — Compressed Air Systems portal
- US DOE — Improving Compressed Air System Performance Sourcebook, Third Edition
- US DOE — Minimize Compressed Air Leaks, Tip Sheet #3
- DEDE — Compressed-air energy-efficiency knowledge hub
- DEDE — Sample Energy Management Manual based on Thai guidance
- DEDE — Compressed-air whole-system automation programme
- PEA — September 2026 electricity tariff structure
- ERC — Ft adjustment history
- ISO 50001:2018 — Energy management systems
*All ROI values in this article are illustrative assumptions. Confirm safety, equipment, tariff, legal, certification and support-programme conditions with the factory’s EHS, maintenance, electrical and finance teams, the OEM, the utility, DEDE and qualified professionals where required.*
