Ultrasound condition monitoring is a condition-based maintenance technique that listens to high-frequency sound (roughly 20 kHz to 100 kHz) produced by friction, turbulence, and electrical discharge to catch faults long before they become audible or visible. Because these ultrasonic signals appear at the earliest stage of a developing problem, ultrasound gives maintenance teams one of the longest lead times of any inspection method. It works on three very different failure modes at once: compressed-air and gas leaks, early-stage bearing and lubrication defects, and electrical arcing, tracking, and corona. This guide explains how each detection mode works, what the numbers look like, and how to fold ultrasound into a data-driven maintenance program.
Human hearing tops out near 20 kHz. Above that, a surprising amount of machine and process activity keeps generating sound that we simply cannot hear. Turbulent airflow through a tiny leak, metal-on-metal contact as a bearing loses its lubricant film, and the ionization of air around a high-voltage defect all emit short-wavelength ultrasound. An ultrasonic instrument uses a sensor and a technique called heterodyning to shift that inaudible band down into a range a technician can hear through headphones and measure as decibels.
Two coupling paths matter. Airborne ultrasound travels through air and is ideal for leaks and electrical discharge you can scan from a distance. Structure-borne ultrasound travels through metal and is captured with a contact probe touching a bearing housing or gearbox. Understanding both paths is what lets one handheld tool cover pneumatics, rotating equipment, and switchgear on a single route.
Compressed air is one of the most expensive utilities on a plant floor, and leaks are the largest hidden loss. A leak creates turbulence, and turbulence radiates broadband ultrasound that a scanner picks up from several meters away, even through background factory noise. Because the technique is directional, an inspector can pinpoint the exact fitting, coupling, or hose that is leaking.
Worked example. Suppose a routine scan finds ten leaks that each measure roughly 3 mm equivalent orifice at 6 bar (about 90 psi). An orifice that size passes on the order of 1 liter per second of free air, so ten leaks waste about 10 L/s. Over a 6,000-hour operating year that is roughly 216,000 cubic meters of compressed air. If generating 1 cubic meter of compressed air costs about 0.02 EUR in electricity, those ten leaks quietly burn about 4,300 EUR per year, and they run whether or not the line is producing anything. A single afternoon of ultrasonic leak detection typically finds enough leaks to pay for the instrument several times over.
The workflow is simple: scan on a fixed route, log each leak location and decibel level, raise a repair task, then re-scan after the fix to confirm the reading dropped. Tracking that loss over time also feeds directly into overall equipment effectiveness conversations, because unstable air pressure from unrepaired leaks causes micro-stops and speed losses on the equipment downstream.
Ultrasound is often the first technology to warn of a failing bearing, earlier than vibration analysis and much earlier than temperature. As a rolling-element bearing begins to degrade, the lubricant film thins and asperities on the metal surfaces start to make contact. That contact generates friction and tiny impacts, both of which show up as a rising ultrasonic decibel level and a change in sound quality through the headphones.
A practical program establishes a baseline dB reading for each bearing when it is known to be healthy. A rise of roughly 8 dB over baseline commonly signals the onset of a lubrication problem, and a rise of about 16 dB or more suggests advanced friction or mechanical damage. Because the decibel scale is logarithmic, an 8 dB increase is not a small nudge; it represents a large jump in emitted energy and is a reliable early trigger.
Ultrasound is also the best available way to confirm correct lubrication. Grease a bearing while listening, and the dB level drops back toward baseline the moment the film is restored. Stop as soon as the sound stops improving. That prevents the classic mistake of over-greasing, which blows out seals and is itself a leading cause of premature bearing death. This ties ultrasound tightly to autonomous maintenance routines run by operators and to a broader shift from reactive to proactive maintenance. Each early catch also improves your MTBF and MTTR reliability metrics by moving repairs off the failure curve.
Electrical faults in switchgear, transformers, and insulators ionize the surrounding air, and that ionization emits ultrasound. Inspectors distinguish three signatures. Corona is a steady, lower-energy discharge that appears above roughly 1,000 volts and sounds like a constant frying hiss. Tracking is intermittent and often follows a contaminated surface, sounding like erratic sputtering. Arcing is a violent discharge with a sharp, popping character and is the most urgent to act on.
Airborne ultrasonic inspection of energized electrical assets is valuable because it can be done from a safe distance and through cabinet louvers without opening panels, reducing arc-flash exposure. It also complements infrared thermography: many discharge faults radiate ultrasound before they generate enough heat for a thermal camera to see, so the two methods together catch more than either alone. Feeding these findings into a structured FMEA helps rank which discharge findings deserve immediate outage versus scheduled repair.
Ultrasound earns its keep when it is a repeatable route, not a one-off hunt. A sound program includes a few disciplines:
Route data is far more useful when it lives in the same system as your other maintenance information. Applying Pareto analysis to logged findings quickly shows which lines and asset classes generate most of your losses, so limited inspection hours go where they pay back fastest. Ultrasound then becomes one input to a wider condition-based maintenance strategy that also includes vibration and thermography.
Ultrasound tells you a fault is starting; you still need a system to act on it and to prove the action worked. That is where Fabrico serves as the real-time data foundation. When a route flags a bearing over threshold or a leak on a compressed-air header, the finding becomes a work order in the Fabrico CMMS, linked to the exact asset, with the technician, spare parts, and preventive schedule all tracked in one place. Fabrico is EU-built with EU data residency, so that maintenance history stays governed.
Fabrico also connects the shop-floor consequences of these faults to production reality through real-time OEE monitoring, including computer vision on machines that have no PLC. An unrepaired air leak that starves a line, or a bearing sliding toward failure, shows up as availability and speed loss in the OEE data, which makes the business case for the repair concrete. Fabrico does not perform the ultrasonic measurement itself and does not claim finished predictive-maintenance modeling; it is the field-ready record and monitoring layer that turns each ultrasound finding into a tracked, verifiable action.
They are complementary rather than competing. Ultrasound usually detects the very earliest friction and lubrication changes, giving the longest warning, while vibration analysis is stronger at diagnosing the specific defect (inner race, outer race, or cage) once damage has developed. Many reliability teams use ultrasound as the low-cost first-alert screen on a wide route, then apply detailed vibration analysis to the assets ultrasound flags.
Basic leak detection and grease-on-sound lubrication can be learned quickly and are often handled by operators as part of their routine. Interpreting subtle bearing sound quality and classifying electrical discharge takes more training and experience. A common approach is to let trained operators cover the simple, high-volume tasks and reserve a certified analyst for the diagnostic calls.
It depends on asset criticality and duty. Compressed-air leak surveys are frequently done quarterly because leaks reappear as fittings vibrate loose. Critical bearings on continuous equipment may be checked monthly or even weekly, while electrical switchgear is often inspected on a scheduled interval aligned with your maintenance calendar. Trending the data over time tells you whether your interval is right.
Ready to turn every ultrasound finding into a tracked work order and see its real impact on production? Book a Fabrico demo to see how real-time OEE and a field-ready CMMS give your condition-monitoring program a single source of truth.
Login
Blog