Infrared Thermography of Electrical Panels: What to Inspect and What the Temperature Rise Means
The short answer
Thermography finds loose and corroded connections before they fail, and the number that matters is not the absolute temperature but the rise above a reference: either the same phase on a similar circuit, or the ambient inside the panel. A widely used scale treats a 1 to 3 degree rise over a similar component as worth monitoring, 4 to 15 degrees as a repair at the next planned outage, and above about 15 degrees as urgent. Above roughly 40 degrees over ambient, treat it as immediate.
Why connections, and why heat
Almost every electrical failure in a panel begins as resistance where there should be none: a terminal that was never torqued, one that loosened through thermal cycling, an aluminium-to-copper joint that oxidised, a crimp made on the wrong die. Resistance times current squared is heat, and heat accelerates the oxidation that created it. The failure is slow, then sudden.
An infrared camera sees that heat through nothing at all, which is the catch below.
The route
Survey under load, ideally above 40 per cent of rated current, and record the load. A panel scanned at 10 per cent load will show you nothing and will produce a clean report that is worthless. This is the single most common reason a thermography programme finds no faults for a year and then has a fire.
Work through, in this order:
- Incomer and main breaker terminations, both cable and busbar sides.
- Busbar joints and droppers, which are the highest-consequence findings.
- Outgoing breaker and contactor terminals, line and load.
- Overload relays and their terminals, which run warm normally, so compare phase to phase.
- Neutral and earth bars, where a hot neutral usually means unbalance or harmonics rather than a loose joint.
- Control transformers, power supplies and drives, which have their own normal operating temperatures.
- Cable glands and terminations at the field end, which are missed because they are not in the panel.
Compare phases first. Three phases carrying a balanced load should read within a degree or two of each other, so the odd one out is the finding even when no absolute number looks alarming.
Four mistakes that produce wrong answers
Scanning through a closed door. Infrared does not pass through a metal panel door or a polycarbonate window that is not IR-transparent. Either open the panel under a permit with the correct PPE and arc-flash assessment, or fit purpose-made IR windows. A survey done through a closed door measures the door.
Ignoring emissivity. Shiny bare copper and aluminium have very low emissivity and read far cooler than they are. Bare busbar can be 30 degrees hotter than the camera reports. Look at the insulation and the terminal body next to the joint, or apply an emissivity target.
Reflected temperature. A shiny surface reflects the surveyor, the lamps and the sun. Move and look again. A hot spot that moves when you move is a reflection.
Comparing a loaded circuit with an unloaded one. A feeder at 80 per cent will always be hotter than one at 5 per cent, and that is not a fault.
Recording a finding usefully
A finding needs the thermal image, a matching visual image, the load at the time, the ambient, the emissivity used, the reference component compared against, and the delta. Without the load, nobody can compare next year's survey with this one, and the programme becomes a folder of pictures.
Where it fits
Thermography is the electrical half of condition monitoring; vibration is the mechanical half. The infrared thermography course covers camera setup, emissivity and reporting, and control panel design covers why the joints were made the way they were, which is what you are inspecting. For the mechanical side, see vibration analysis.
Frequently asked questions
How often should panels be surveyed? Annually for general distribution, six-monthly for critical or heavily cycled switchgear, and always after any work that disturbed terminations.
Does a torque check replace thermography? No, and neither replaces the other. Torque proves the joint was tight when you checked it; thermography finds the joints nobody checked and the ones that loosened since.
Can maintenance staff do this, or is a certified thermographer needed? Staff can run a screening route usefully after basic training. A Level I or II thermographer is worth having for interpretation, reporting and anything that will be used to justify an outage.
