Most electrical repairs end administratively, not technically: the work order closes, the invoice gets paid, and the original finding is assumed dead because money was spent on it. For low-severity items that is a reasonable economy. For the findings that justified an urgent repair — the ones rated serious because a joint was running tens of degrees above its neighbours — assumption is the wrong tool, because the failure that follows an unverified repair costs exactly what the original failure would have.
Why the fault can survive its own repair.
A hotspot is a symptom: something in the current path has developed resistance, and the current turns that resistance into heat. The repair usually attacks the obvious cause — a loose connection gets retorqued, a corroded lug gets replaced, a load gets rebalanced. But heat damages the material around it. Copper annealed by months of overheating goes soft and stops holding the clamping force it was torqued to; a contact surface pitted and oxidised by heat makes a poor connection even at the right torque; a replaced lug crimped onto heat-damaged conductor starts the same cycle again. The work was done, the invoice is honest — and the fault is still there.
That is why the question "was the work done?" and the question "is the fault gone?" have different answers, and different evidence.
The confirmation standard.
Confirmation is a re-scan that repeats the conditions of the original finding closely enough that the two images can be honestly compared. Four elements make it defensible:
Comparable load, recorded both times
Heat rises with current. The original report should state the load at the time of the finding; the re-scan should run at something comparable — and record it. An after-image at half the load proves nothing except that half the load makes less heat.
The same point, findable
The re-scan targets the exact terminal or joint from the finding — which is why the original report needed the thermal and visual image pair, labelled. You cannot verify a repair you cannot locate.
The same comparison logic
The finding was rated on a temperature difference against a stated reference — the same phase on a neighbouring circuit, or ambient. The verification uses the same reference and the same severity bands, so "resolved" means the same thing "Danger" meant. Our Shah Alam case study shows those bands printed in full.
Before and after, side by side
The closing evidence is the pair: original image and re-scan image, loads and references stated, with the rating change written down and a name behind it. That page is what an auditor, an insurer, or your own successor can rely on.
The four substitutes to refuse.
The visual check. "Terminals inspected, connections clean" describes what the eye can see, and resistance is invisible to the eye — the whole reason the survey used a thermal camera in the first place. The torque check. Correct torque is necessary and insufficient: it confirms the wrench setting, not the condition of the metal underneath it. The quiet-hours scan. A re-scan taken during a shutdown, at idle, or on a weekend reads cool because everything reads cool — a joint that carries no current makes no heat. The absolute number. "It reads 45°C now" is not a verdict; without the load and the reference, a single temperature is trivia. The original finding was a difference under stated conditions, and so is its closure.
A fifth item is not a substitute but a structural caution: when the surveyor, the repair contractor and the verifier are the same company, the re-scan is marking its own homework. Usually that is still done honestly — but on a large repair bill, the before/after pack is exactly the document worth an independent pair of eyes, and we have written separately about reading the report before approving the repairs. The same five checks apply to the closure evidence.
Verify twice, then let the trend carry it.
The first verification happens soon after the repair, once normal load has been restored — days, not months, for a finding that was rated serious. The second is quieter: the point stays on the survey list, and the next scheduled scan confirms the repair has held under months of real operation. A joint that was repaired properly is boring in every subsequent image; a joint that drifts warm again is telling you the damage ran deeper than the fix.
That second look is where one-off inspections quietly fall short — there is no "next image" to compare against unless someone schedules one. It is also, declared plainly, one of the reasons we sell assessment as a year-on-year programme: repaired points stay mapped, and every later survey re-reads them against the same baseline by default.
What should end up in the file.
One page per closed finding: the before image and the after image, the load at both scans, the reference and severity logic used, the repair action taken, the new rating, the date, and the name of the person who verified it. If a finding was closed without that page, it was not verified — it was archived. The difference matters on the day the insurer, the auditor, or the fire investigator asks to see the file.
Retrieved from https://amkatechnologies.com/insights/verify-electrical-hotspot-repair
Published 30 July 2026. AMKA Technologies Sdn Bhd, SSM 202301041763 (1535682-T).