Key takeaways
Short answer: Ultrasonic testing and thermography are two condition-monitoring techniques that sense different physical symptoms of a fault. Ultrasonic testing detects high-frequency sound — beyond human hearing — produced by friction (early bearing wear), turbulence (gas and vacuum leaks), and electrical discharge (arcing, corona). Thermography uses an infrared camera to detect heat — the temperature differences that reveal loose electrical connections, overloaded circuits, friction, and blocked flow. Ultrasound often gives the earliest warning, because friction and electrical discharge emit sound before they generate significant heat; thermography excels at electrical inspection and scanning wide areas quickly. They are complementary, not competing — and both sit alongside vibration analysis in a condition-monitoring program.
In condition monitoring, ultrasonic testing means detecting and analysing high-frequency sound — typically above 20 kHz, beyond human hearing — that machinery and systems emit when something is wrong. An ultrasound instrument translates this inaudible sound down into the audible range and into a measurable level. Three families of fault generate it. Friction: a bearing starting to lose lubrication or wear produces a rising ultrasonic signature long before it gets hot or audibly rough, making ultrasound one of the earliest warnings of bearing trouble and a precise guide for lubrication. Turbulence: gas, air, steam, and vacuum leaks create broadband ultrasound at the leak point, so a technician can pinpoint compressed-air leaks (a major energy cost) and valve passing. Electrical discharge: arcing, tracking, and corona in switchgear and high-voltage equipment emit ultrasound, allowing detection of electrical faults from a safe distance. (This is distinct from ultrasonic testing used as an NDT method to measure wall thickness or find internal cracks — same physics, different application.) Ultrasound's strength is catching faults very early and locating them precisely.
Thermography, or infrared thermography, uses a thermal imaging camera to visualise the heat that surfaces emit, turning temperature differences into an image. Because almost every developing fault eventually produces abnormal heat, thermography is broadly useful. In electrical systems it is the workhorse for finding loose or corroded connections, overloaded conductors, and imbalanced loads — all of which run hotter than their healthy neighbours, showing up as bright spots in the image. In mechanical systems it reveals overheating bearings and couplings, misalignment, and friction. In process equipment it finds blocked or fouled flow, failed steam traps, refractory and insulation breakdown, and liquid levels in vessels. Its great advantages are that it is non-contact, can scan a wide area quickly (a whole switchboard or motor line in one sweep), and produces an intuitive image that is easy to interpret and document. The trade-off is that heat is often a relatively late symptom: by the time a fault is clearly hot, it may be further along the failure path than an ultrasound or vibration signature would have revealed.
The core difference is the physical symptom each senses: ultrasound listens, thermography looks. Ultrasonic testing detects the high-frequency sound of friction, turbulence, and electrical discharge; thermography detects the heat those and other faults produce. This makes them sensitive to overlapping but not identical fault sets. Some problems announce themselves loudly in ultrasound but barely in heat — a small compressed-air leak makes plenty of ultrasonic noise but negligible temperature change, and early bearing friction is ultrasonically obvious while still thermally quiet. Others are clearer in heat — a loose electrical connection under moderate load may show a strong thermal signature, and an overloaded circuit is fundamentally a heat phenomenon. Many serious faults eventually produce both: a degrading bearing first emits ultrasound, then vibration, then finally heat. Because the two techniques key off different physics, they catch different things and confirm each other when they overlap, which is exactly why mature programs run both rather than choosing one.
On the P-F curve — the interval between a fault becoming detectable (P) and functional failure (F) — ultrasound and thermography typically sit at different points. For friction-driven mechanical faults and for electrical discharge, ultrasound generally detects earlier: friction generates high-frequency sound before it generates meaningful heat, so an ultrasonic warning can precede a thermal one by weeks. Thermography tends to catch the same mechanical fault later, once enough heat has built to stand out — though for purely electrical-resistance faults (a loose connection heating under load), thermography can be the most direct and earliest indicator. The practical implication is sequencing: ultrasound (and vibration) for the earliest possible warning on rotating equipment, thermography to confirm and to own the electrical-connection and surface-temperature domain. Earlier detection means a longer P-F interval, which means more time to plan the repair rather than react to a failure — the whole economic point of condition monitoring.
Take a motor-driven pump on a monthly monitoring route. The bearing begins to fail. Ultrasound flags it first: at roughly four weeks out, the ultrasonic friction level on the bearing has climbed well above its baseline, prompting a lubrication check and a planned bearing replacement. Thermography on the same bearing shows nothing remarkable until about one week before failure, when it finally runs hot — useful confirmation, but a much shorter runway. Now take two other findings on the same route. Ultrasound locates a compressed-air leak on a nearby header that thermography cannot see at all, because the leak barely changes temperature — yet it is wasting energy continuously. Thermography, meanwhile, finds an overloaded breaker in the motor-control centre running 30 degrees hotter than its neighbours — a developing electrical fault that ultrasound, with no arcing yet, would not have flagged. The example shows the pattern: ultrasound caught the early mechanical and the leak; thermography caught the electrical heat. Each found something the other missed, and together they covered far more than either alone.
Use ultrasonic testing when you want the earliest warning of friction and bearing faults, when you need to find and quantify compressed-air, gas, steam, or vacuum leaks, when checking for valve passing or steam-trap function, and for detecting electrical discharge (arcing, corona, tracking) in switchgear from a safe distance. Use thermography when inspecting electrical connections and loads for hot spots, when you need to scan a large area or many components quickly, for surface-temperature problems like overheating, blocked flow, failed steam traps, and insulation or refractory breakdown, and where a documented thermal image is valuable evidence. The honest answer for most reliability programs is not "which one" but "both," deployed where each is strongest, and combined with vibration analysis for rotating equipment. The techniques overlap enough to cross-confirm and differ enough to cover each other's blind spots; choosing only one leaves a category of faults unmonitored.
Both techniques protect the Availability factor of OEE by catching developing faults before they cause unplanned breakdowns. Every bearing caught by ultrasound weeks early, and every overheating connection caught by thermography before it fails, is a breakdown that does not happen — downtime avoided and Availability preserved. The earlier the detection, the longer the planning window, so the repair can be scheduled into a low-impact slot rather than forcing an emergency stop. This is the same reliability logic behind availability and reliability: fewer unexpected failures means higher uptime. Condition monitoring also supports a condition-based maintenance strategy, where you act on the actual evidence of degradation these techniques provide rather than on a fixed calendar — doing maintenance when the data says it is needed, which both prevents failures and avoids unnecessary intervention.
Condition-monitoring techniques tell you a fault is developing; Fabrico connects that to the production cost of doing nothing. By tracking downtime causes and their effect on live OEE, it shows which equipment is actually draining availability — helping you target ultrasound and thermography routes where they will protect the most output, and confirming that catching faults early actually reduced the unplanned breakdown losses in the OEE trend. Book a demo to link condition monitoring to real availability gains.
Ultrasonic testing detects high-frequency sound from friction, leaks, and electrical discharge; thermography detects heat with an infrared camera. Ultrasound listens for the earliest mechanical and leak symptoms; thermography sees temperature problems, especially electrical hot spots. They sense different physical symptoms and are complementary.
For friction, bearing wear, and leaks, ultrasound usually detects earlier because these produce high-frequency sound before significant heat. For loose electrical connections heating under load, thermography can be the earliest and most direct indicator. Using both gives the widest early-warning coverage.
Generally no — a compressed-air leak produces a lot of ultrasound but almost no temperature change, so it is largely invisible to thermography. Ultrasonic testing is the standard tool for locating and quantifying compressed-air and gas leaks.
For most reliability programs, both. Use ultrasound for early friction, leaks, and electrical discharge; use thermography for electrical connections, surface temperature, and fast wide-area scanning. They cover each other's blind spots, and many programs add vibration analysis for rotating equipment.
By catching developing faults before they cause unplanned breakdowns, they protect the Availability factor of OEE. Early detection lengthens the planning window so repairs can be scheduled into low-impact slots instead of forcing emergency stops, reducing downtime losses.
Anmeldung
Blog
