Almost every electrical failure passes through a stage where something gets hot before it breaks. A joint that has loosened, a phase carrying more current than its neighbours, or a breaker with a failing contact all shed heat that is invisible to the eye but obvious to an infrared camera. Electrical thermography inspection is the practice of surveying energized panels and connections with such a camera to catch those hot spots while the equipment is still running. This guide explains what a thermal survey looks for, how delta-T severity criteria turn a warm image into a ranked repair, and how the results become maintenance work orders.
Electrical thermography in one line: Electrical thermography inspection is a condition-monitoring survey in which an infrared camera images energized electrical equipment to find abnormal heat caused by loose connections, overloaded or imbalanced phases, and failing components. Because a developing fault heats up before it fails, the survey catches problems early, and the temperature rise of each hot spot, its delta-T, is used to rank findings by urgency so repairs can be scheduled before the fault causes an outage.
The most common find in an electrical thermal survey is the high-resistance connection. When a bolted lug, breaker terminal, or fuse clip loosens or corrodes, its contact resistance rises, and because power dissipated at a joint climbs with the square of current, even a small resistance increase turns into real heat under load. On a thermal image this shows up as a bright, localised hot spot right at the connection while the conductor either side runs cooler. These are the bread and butter of electrical thermography because they are common, they progress toward failure, and they are cheap to fix once found.
The second big category is load-related heating. Comparing the three phases of a circuit reveals imbalance: if one phase runs noticeably hotter than the other two along its whole length rather than at a single joint, it is likely carrying more current, pointing to an unbalanced load, a harmonics problem, or a distribution fault upstream. A whole component running uniformly hot, such as a breaker, contactor, or transformer, can indicate overloading or an internal problem rather than a bad connection. Reading whether the heat is at a point or spread out is central to diagnosing what is actually wrong.
For the survey to be valid the equipment must be energized and, ideally, carrying a representative load, because a fault that only heats under current will be cold on an idle circuit. Inspectors note the load at the time of the reading, because a hot spot found at half load will be considerably worse at full load. They also account for surface emissivity and reflections, since shiny bus and painted surfaces radiate differently, and they open covers safely or use inspection windows so the camera sees the actual connections rather than a warm enclosure wall.
A thermal image on its own is just a coloured picture; it becomes actionable when the anomaly is quantified. The standard way to do that is delta-T, the temperature difference between the hot component and a reference. That reference is usually a similar component under similar load, such as the same terminal on another phase, or the ambient air where no comparable component exists. Measuring the difference rather than the absolute temperature is what makes the reading meaningful, because a joint at sixty degrees is unremarkable on a hot day and alarming on a cold one, whereas a thirty-degree rise above its neighbours is a problem in either case.
Severity criteria then map delta-T ranges to repair urgency. Published guidance, such as the tables associated with NETA maintenance testing, groups anomalies into bands: a small rise might be logged and rechecked at the next survey, a moderate rise scheduled for repair at the next convenient shutdown, and a large rise flagged for prompt or immediate action. The exact numbers differ between the phase-to-phase comparison and the phase-to-ambient comparison, and a facility usually adopts one recognised scheme so every inspector ranks findings the same way. The point is that two inspectors looking at the same hot spot should assign the same priority.
Turning that ranking into a good decision still needs judgement about load. Because heating scales with load, a delta-T measured at partial load understates the real severity, so inspectors either normalise the reading toward expected full load or explicitly note the load so the reviewer can weigh it. The output for each finding is a package of thermal image, visual photo, location, measured delta-T, reference basis, load, and an assigned severity band. That package is what makes a thermographic result defensible and repeatable rather than a matter of one person's impression.
A thermal survey is only useful if its findings turn into work that actually gets done. Each ranked anomaly should become a work-order candidate in the maintenance system, carrying its severity so the planning process treats a large delta-T differently from a small one. High-severity findings drive prompt corrective work orders, moderate findings are bundled into the next planned shutdown, and low-severity findings are logged to be re-checked at the following survey so their trend can be watched. Without this hand-off the survey becomes a report that sits in a drawer while the fault it found keeps getting hotter.
Trending matters as much as the one-time reading. A connection that measured a small rise last year and a larger rise this year is deteriorating, and comparing successive surveys of the same asset reveals that trajectory in a way a single snapshot cannot. That is why thermography findings belong in the asset's ongoing condition history alongside its other data, so a reviewer can see not just today's delta-T but whether it is climbing survey over survey. A hot spot that is stable is a different problem from one that is accelerating.
This is where a central monitoring platform earns its place. In oil and gas and other distributed operations, electrical gear lives in switchrooms scattered across many remote facilities, and a cloud SCADA platform such as Merobix lets thermographic findings be logged against the specific asset that produced them, ranked, trended across surveys, and turned into work orders in the same place operators already watch live current, load, and breaker status. Correlating a phase-imbalance finding with the live load data the platform is already collecting sharpens the diagnosis, and keeping the whole electrical health record with the asset means a hot connection found on a routine survey is not lost between the field and the office.
Yes. Most electrical faults only generate heat when current is flowing, so the equipment must be energized and ideally carrying a representative load. A loose connection or imbalanced phase found at full load can be missed entirely on an idle circuit. Inspectors record the load at the time of each reading because a hot spot found at partial load will be worse at full load.
Delta-T, the temperature difference from a similar component or from ambient, removes the effect of the surroundings so the reading reflects the fault itself. A connection at sixty degrees means little without knowing the ambient and load, but a thirty-degree rise above an identical component on another phase is clearly abnormal. This makes findings comparable across surveys, seasons, and sites.
Each anomaly is quantified with a delta-T, assigned a severity band using a recognised criteria table, and turned into a work-order candidate in the maintenance system. High-severity findings drive prompt repairs, moderate ones are scheduled for the next shutdown, and minor ones are logged to be re-checked next survey so their trend can be watched. The severity ranking is what lets planners prioritise a limited repair budget on the connections most likely to fail.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.