A leaking fitting or a valve that will not seal makes a sound, but most of that sound is at frequencies far above human hearing. Airborne ultrasound leak detection listens in that ultrasonic range, where turbulent gas escaping a small opening broadcasts a broadband hiss that stands out sharply against a quiet background. A handheld instrument then shifts those inaudible signals down into the range an inspector can hear and localize. It is one of the most effective ways to find compressed-air waste and valve leakage that pressure gauges and trending never reveal.
Airborne ultrasound in one line: Airborne ultrasound leak detection finds pressurized gas and air leaks by detecting the broadband ultrasound produced when escaping flow becomes turbulent at the leak orifice. A directional ultrasonic receiver, typically tuned around 40 kHz, uses a heterodyne circuit to shift the ultrasound down into the audible range so an inspector can hear the leak and pinpoint its location.
When a gas under pressure escapes through a small opening, it accelerates and the smooth flow breaks up into turbulence right at the orifice. Turbulence is chaotic, small-scale swirling that generates sound across a very wide band of frequencies, and a large share of that energy falls in the ultrasonic range above the limit of human hearing. This is the physical basis of the method: the leak itself is an ultrasound source, and the more turbulent the escaping flow, the stronger and broader the ultrasonic emission. The sound is generated at the leak point, which is why the emission can be traced back to the exact spot.
Two properties of this ultrasound make it ideal for detection. First, it is highly directional and short-range in air; high-frequency sound does not travel far or bend around obstacles the way low-frequency sound does. That means an inspector can sweep a directional receiver across equipment and hear the signal spike sharply only when the sensor points straight at the leak, giving precise localization. Second, ordinary industrial background noise, motors, flow, machinery, is dominated by lower frequencies, so the plant is comparatively quiet in the ultrasonic band. A leak's broadband ultrasound therefore stands out against a low-noise background even in an environment that is deafening to the human ear.
The method detects both gas escaping outward and, in the case of valves, gas passing through internally. A pressurized leak to atmosphere makes ultrasound at the exit; a valve that is supposed to be shut but is passing makes ultrasound as the gas rushes turbulently through the throttled gap inside the valve. That internal turbulence radiates through the valve body and can be picked up by holding the sensor against the valve, which is how ultrasound finds leaking or passing valves that show no external sign at all.
The receiver typically centers on around 40 kHz, well above the roughly 20 kHz upper limit of human hearing, so on its own the signal is inaudible. The instrument makes it usable through heterodyning, the same frequency-shifting principle used in radio. The incoming ultrasound is mixed with an internal reference frequency, and the mixing produces a difference frequency that falls within the audible range. What was a 40 kHz signal the inspector could never hear becomes a hiss or rushing sound in the headphones, preserving the character and the changing intensity of the leak so the ear can follow it.
This audible translation is what makes the tool so intuitive to use. As the inspector aims the receiver and moves closer to the source, the pitch and loudness of the shifted signal rise, and they fall away as the sensor turns off-target. Homing in on a leak becomes a matter of following the sound to its peak, exactly the skill people already have for locating a noise. Alongside the audio, instruments usually display an intensity reading so the inspector can compare leaks quantitatively and record how loud each one is, giving a repeatable measure rather than only a subjective impression.
Accessories extend where the method reaches. A narrow probe or a rubber focusing cone concentrates the sensor on a single fitting in a crowded area, and a parabolic reflector or dish gives the range to scan overhead pipe racks and elevated lines from the ground. Because the underlying signal is directional ultrasound, these simple acoustic accessories meaningfully change the sensitivity and reach of the same instrument, letting one tool cover everything from a tight manifold to a distant overhead run.
The reason ultrasound earns its place is that many leaks are invisible to the process instrumentation. A compressed-air system riddled with small leaks may hold near-normal header pressure because the compressors simply run more to make up the loss; the pressure trend looks fine while energy quietly drains away. A valve that is passing slightly does not necessarily move any gauge an operator watches. Pressure and flow trending see the aggregate effect at best and often nothing at all, because no single small leak is large enough to register against the noise of normal operation. Ultrasound goes straight to the individual leak and localizes it.
This complements rather than competes with a SCADA monitoring program. A cloud SCADA platform such as Merobix is excellent at the system view, trending header pressure, compressor run time and loading, and total flow, and it can reveal that a system is losing more air than it should or that make-up compression is creeping up over time. That aggregate signal is exactly what tells a plant it has a leak problem worth chasing. What the trend cannot do is point at which fitting or valve is responsible; an ultrasound survey then walks the system and finds the specific culprits behind the number the platform flagged.
Used together, the two form a full loss-management loop. The SCADA history quantifies the problem and, just as importantly, verifies the fix: after leaks found by ultrasound are repaired, the platform should show make-up compression and total flow drop back, confirming the survey actually recovered the loss. Trending those metrics over time also catches the reappearance of leaks so surveys can be scheduled when the numbers justify them rather than on a fixed calendar. The ultrasound instrument finds and localizes; the SCADA trend measures the scale of the problem and proves the recovery, and for passing valves the same pairing links a specific leaking valve to the throughput or efficiency loss the historian records.
When gas under pressure escapes through a small opening it accelerates and the flow becomes turbulent right at the orifice. Turbulence generates sound across a very wide band of frequencies, and much of that energy falls in the ultrasonic range above human hearing. The leak point itself is the ultrasound source, so the emission can be traced back to the exact spot, and stronger turbulence produces a stronger, broader ultrasonic signal.
The ultrasound from a leak, centered around 40 kHz, is far above the roughly 20 kHz limit of human hearing, so it cannot be heard directly. Heterodyning mixes the incoming ultrasound with an internal reference frequency to produce a lower difference frequency that falls in the audible range. This lets the inspector hear the leak as a hiss in headphones, following its loudness and pitch to pinpoint the source.
Small leaks often do not move the process instrumentation, because compressors simply run more to hold header pressure and a slightly passing valve may not shift any gauge. Pressure and flow trending see only the aggregate effect, or nothing, since no single small leak is large enough to register. Ultrasound detects and localizes each individual leak directly, so it finds the specific fittings and valves behind a loss that the system-level trend can only hint at.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.