Most compressed-air leak surveys do not fail during detection. They fail after the report is issued. Photographs are saved, tags fall off, production cannot release the line, and nobody checks whether repaired items actually stopped leaking. Six months later the plant orders another survey and finds many of the same faults. A survey that “sticks” must be designed as a maintenance process from the beginning.
For plants in Pakistan, the practical constraints are familiar: several shifts, limited shutdown windows, overhead pipework, restricted process areas, and a mixed network that has grown over years. At Jams (Pvt.) Ltd, we use a five-stage structure—baseline, find, prioritise, repair, verify—so the technical work can move through the plant’s existing responsibilities.
1. Define the boundary and baseline
Draw the survey boundary before walking the line. Identify compressors, dryers, receivers, main headers, pressure zones, departments, and normally open cross-connections. Record the shift plan and equipment that must remain pressurised outside production. If the network is not understood, a night-flow number can easily be mistaken for leakage when it includes dryer purge, automatic drains, instrumentation, or essential users.
Where suitable metering exists, capture at least one representative production profile and one defined non-production window. A VA 500 flow meter at the main header or a correctly selected branch meter can establish demand before the campaign. Also record pressure. A lower baseline after repairs is encouraging only if the compared periods had similar pressure, connected loads, and production state.
2. Prepare a route that can be completed safely
Divide the site into zones that match maintenance ownership. A route should include mains, drops, flexible hoses, filter-regulator assemblies, valve islands, cylinders, quick couplings, cabinet air, pneumatic tools, and abandoned branches. Add high-level or restricted assets to a separate access plan rather than letting surveyors improvise around ladders or moving equipment.
Agree the survey conditions with production. Some leaks appear only when a cylinder is extended, a valve is energised, or a machine is idle. The person accompanying the survey should understand those states and have authority to operate equipment safely. Detection never overrides lockout, permit, hot-work, hazardous-area, or access controls.
3. Select tools for the plant environment
A conventional ultrasonic detector is useful for close work and source confirmation. Acoustic imaging is valuable for broad, noisy, or elevated areas where several sources may be present. The LeakCam 600 uses a microphone array and beamforming to visualise ultrasonic sources, allowing the surveyor to scan an area and then investigate the indicated component.
Neither method should be treated as automatic truth. Reflections, pneumatic exhausts, vacuum equipment, and other ultrasonic sources can confuse interpretation. Re-scan from another angle, correlate the source with the component, and observe whether it changes with machine state.
4. Use tags that lead a fitter to the exact point
Every finding needs a unique ID shared by the physical tag, image, report, and maintenance work order. The record should contain enough information for someone who was not on the survey:
- Plant, building, level, department, machine, and component.
- Exact joint or part, with a wide location image and a close identification image where possible.
- Date, time, pressure zone, and relevant machine state.
- Estimated severity or loss basis, clearly labelled as an estimate.
- Safety access, isolation, scaffold, lifting, or shutdown requirement.
- Likely repair type and the team responsible for approval and execution.
Use tags suited to heat, oil, dust, and weather at the location. A handwritten paper label tied to an outdoor header is not a reliable link to the work order. Where plant rules permit QR or barcode labels, they can reduce transcription, but the visible ID should remain readable without a device.
5. Prioritise a repairable backlog
Estimated loss is one input, not the only ranking criterion. Consider annual pressurised hours, operating pressure, process consequence, safety, access cost, required outage, and recurrence. Quick items can be grouped into a short maintenance window. Larger valves, corroded pipe, or inaccessible headers may need an engineered work pack.
Assign every item one status: open, planned, repaired-awaiting-verification, verified, deferred-with-reason, or rejected-after-confirmation. “Sent to maintenance” is not a status that supports management. A named owner and target date are more valuable than a long list sorted only by theoretical rupees.
Estimate honestly
Leak instruments and reporting tools may estimate flow from acoustic information and entered pressure. Such estimates help rank work, but uncertainty should be visible. Actual loss changes with pressure and component condition. Annual cost further depends on pressurised hours, compressor specific power, electricity tariff, and system efficiency.
State the assumptions in the report. Do not add every estimate and present the total as guaranteed savings. Some leaks cannot be eliminated immediately; some overlap with non-productive demand measured at the header; and compressor energy does not always fall linearly if controls cannot unload or sequence efficiently.
Repair with root cause in mind
Tightening a fitting may close the current leak but not the reason it returned. Check vibration, unsupported tubing, side load, incorrect thread form, damaged sealing surfaces, unsuitable hose, excessive pressure, heat, chemical exposure, and repeated operator movement. A frequently replaced coupling may need relocation or a different connection concept.
Use approved components and isolation methods. Temporary sealing compounds and improvised clamps can create safety and reliability issues. Update drawings when dead legs are removed or branch arrangements change, because the next survey and the next isolation plan depend on that information.
Verify at two levels
First, re-check the repaired component under the same relevant pressure and machine state. Mark it verified only when the indication is gone or the remaining source is understood. Second, compare zone or header flow over equivalent operating windows. Component verification proves individual closure; flow trending tests whether the campaign changed system demand.
If baseline flow does not fall as expected, investigate instead of forcing a savings narrative. Production may have increased, pressure may be higher, a new consumer may be connected, dryer purge may have changed, or unrecorded leaks may remain. A measurement programme should make those explanations visible.
Set the next survey before closing this one
New leaks begin as soon as equipment returns to service. Use the first campaign to define frequency by risk: high-pressure, high-use, vibration-prone areas may need frequent checks; stable low-use zones may need less. Add operator reporting between formal surveys and re-check modified machines after commissioning.
A short monthly review can track open count, verified closure rate, age of high-priority items, non-production baseline, and recurring component types. These indicators reward actual repairs rather than the number of leaks found.
Start with the information JAMS needs
To scope a survey, send the compressor and dryer arrangement, approximate pipe network, pressure levels, operating shifts, site access rules, and any existing flow data. Use the JAMS contact page to discuss whether the work needs acoustic imaging, close ultrasonic confirmation, temporary or fixed metering, and a verification visit. The deliverable should fit your maintenance system, not require a parallel process nobody owns.
Frequently asked questions
How long should a leak survey take?
Area, access, asset density, operating states, and documentation depth determine duration. A count of compressors or floor area alone is insufficient. A pre-survey route review allows realistic planning and identifies zones that need separate permits or production windows.
Should we shut down the plant for detection?
Leaks generally need the network to be pressurised, and operating surveys reveal sources in normal states. A shutdown or isolation may still be needed for safe access, confirmation, or repair. Coordinate each stage with production and site safety procedures.
What is a good survey KPI?
Verified closure rate and change in comparable non-production demand are stronger than “leaks found”. Also track high-priority backlog age and recurrence. These measures show whether detection is producing maintained results.
When can savings be claimed?
After repair and verification, using declared assumptions or measured before-and-after flow under comparable conditions. Energy conversion should use an agreed compressor performance basis and tariff. Keep estimated potential separate from realised, validated improvement.
