Ultrasonic leak detection is a condition monitoring technique that uses handheld ultrasonic detectors to find leaks and other faults by listening for the high-frequency sound they produce, above the range of human hearing.
When a compressed gas escapes through an orifice, a cracked fitting, or a worn valve seat, the flow becomes turbulent at the point of restriction, generating broadband acoustic energy that extends well above what a person can hear.
That energy is strong and highly directional at close range, so it is easier to pinpoint than the faint hiss a person might catch near a large leak.
Small leaks, leaks in noisy areas, and leaks behind guards or up in pipe racks are usually inaudible to the ear but still radiate a clean ultrasonic signature.
A handheld detector uses a directional sensor to pick up that energy, then shifts it into the audible range so a technician can hear it through headphones while watching a signal-strength display.
Ultrasound attenuates quickly through air and solid barriers, so the signal drops off fast with distance, letting an operator home in on the exact leak point.
Compressed air is one of the most expensive utilities in a plant per unit of energy delivered, and unmanaged leakage is consistently one of the largest sources of waste in a compressed air system. In plants with no leak management programme, it is common for a substantial share of total generated air, often cited in the range of 20 to 30 percent of compressor output, to be lost through leaks rather than reaching productive use. A structured survey, tag, and repair programme is the standard fix: leaks are located, tagged, estimated for flow, logged, and repaired largest-first, with the survey repeated quarterly to twice a year as new leaks develop. The US DOE calls 5 to 10 percent of total system flow a reasonable, cost-effective leak target that is typical for industrial facilities.
| Leak orifice diameter | Approximate air loss at 6 to 7 bar | Typical priority |
|---|---|---|
| 1 mm | Low, about 1 L/s | Adds up across many leaks |
| 3 mm | Moderate, about 10 L/s | Worth prompt repair |
| 6 mm | High, about 40 to 45 L/s | High priority repair |
| 10 mm | Very high, about 110 to 125 L/s | Equivalent to extra compressor capacity |
Figures are for a clean round hole with no flow losses; a sharp-edged crack passes roughly 60 percent of these values. Exact flow depends on line pressure and orifice shape, so treat this as an order-of-magnitude guide, not a substitute for a specific detector reading.
Logging leak tags and trending leak volume over successive surveys is exactly the kind of recurring maintenance loop a CMMS platform like Fabrico is built to support. Book a Fabrico demo to see how leak tags fit into a broader condition monitoring workflow.
Steam trap testing with an ultrasonic detector, often paired with a contact temperature probe, lets an inspector classify a trap as normal, failed open, or failed closed without opening it.
A failed-open trap blows live steam continuously and produces a strong, steady signal, while a failed-closed trap produces little sound and runs cooler than expected upstream. The US DOE recommends testing steam traps weekly to monthly at 150 psig and above, monthly to quarterly from 30 to 150 psig, and annually below 30 psig. A regular trap survey pairs with a wider steam and condensate return inspection programme.
Vacuum leak detection works on the same principle in reverse, locating leaks in packaging lines and vacuum furnaces without a tracer gas, though a helium leak detector remains more sensitive for very fine leaks, with helium sprayed on the outside while the evacuated system is connected to the detector.
Detectors range from simple screening units to advanced models with signal processing, leak-rate calculators, and data logging. The availability of contact probes for solid-borne ultrasound (bearings, steam lines) versus airborne sensors (leaks, electrical faults) is a key selection point, and results depend heavily on operator technique, so training matters as much as specification.
Surveys complement rather than replace other methods; a detector that flags an unusual bearing sound should trigger a proper vibration analysis, and following a recognized framework such as ISO 18436 condition monitoring certification guidance keeps inspection integrated with the reliability programme.
Set the detector to 40 kHz for air leaks, the setting UE Systems gives in its leak survey procedure, and start at the highest sensitivity. Work gross to fine: scan the area, adjust the sensitivity to pinpoint the leak, then fit the rubber focusing probe to cut out competing ultrasound and confirm the exact point by scanning around it in every direction and sealing the probe against it. For a dB reading you can compare and turn into a flow estimate, UE's procedure takes it 12 to 15 inches (about 30 to 38 cm) from the tip of the probe.
The US DOE costs leaks with a simple formula: annual cost = number of leaks x leak rate (cfm) x kW per cfm x annual operating hours x electricity price per kWh. DOE assumes about 18 kW per 100 cfm of compressed air generation and multiplies table flows by 0.61 for sharp-edged holes. In metric terms, Atlas Copco puts a 3 mm hole at 7 bar at 11.1 L/s and 4.0 kW. At 6,000 hours a year and 0.15 per kWh, that is up to 3,600 a year in your currency for one hole, or about 2,200 once the 0.61 sharp-edge factor is applied.
End each survey with a report that shows leaks found, leaks repaired, flow lost and saved, money saved, recommendations, and the next survey date. Retest every repaired leak to confirm the repair worked and that no new leak was created during it. Then check pressures: where leaks have been tolerated for a long time, pressure in parts of the plant has often been raised to compensate, so bring it back down to the right level once the repairs are done.
Yes. The technique listens above normal industrial noise, and the detector is highly directional at close range, so it works on a loud production floor, and where blow-offs or other leaks compete, a rubber focusing probe or shielding isolates the leak.
Most facilities run a full survey two to four times a year, since new leaks develop continuously from vibration and wear at fittings.
Detectors can estimate leak flow and cost from signal strength, distance, and line pressure, accurate enough for prioritizing repairs. A precise measurement needs the leak isolated and measured directly.
Yes, including steam trap testing, electrical corona and arcing, and early acoustic signs of bearing wear, often before those faults show up in vibration or infrared monitoring.
Programați o întâlnire individuală cu experții noștri sau înscrieți-vă direct în planul nostru gratuit.
Nu este nevoie de card de credit!
Login
Blog