The default annual thermal survey is a calendar decision dressed up as an engineering one. Here is a better basis.
Ask most facilities how often their electrical assets receive a thermal imaging inspection and the answer is annual. Ask why annual, and the answer is usually that it has always been annual, or that the insurer or an auditor expects to see a yearly certificate. Neither is an engineering answer. Annual-by-default is a calendar decision wearing an engineering costume, and on some assets it is too often, on the ones that matter most it is nowhere near often enough.
Why one interval for everything is wrong
A thermal survey is a sample. It captures the condition of an asset at one moment, under one load, in one set of conditions. The right question is how quickly a given asset can move from healthy to dangerous, because that rate is what should set the gap between samples. Assets do not degrade at the same rate, so a single interval applied across a whole site guarantees you are over-inspecting the benign and under-inspecting the critical.
Three factors dominate that rate of change, and none of them are the same across a site.
Criticality. An asset tagged essential to operation, feeding equipment the facility cannot afford to lose, carries a consequence of failure that a non-essential lighting circuit does not. Same fault, very different cost, and the inspection interval should reflect the cost, not just the probability.
Load profile. A device that spends its life near the top of its rated capacity has little margin and ages faster under thermal stress. A lightly loaded device on the same board has room to spare. Their sensible inspection intervals are not the same.
History and environment. An asset with a clean, well-documented inspection history behind it earns a longer leash than one that has never been opened, or one that lives in a hot, dirty, or humid environment that accelerates connection degradation.
The trap of the clean scan
Here is where risk-based intervals get misused. A clean survey is often read as permission to extend the interval or drop the asset down the priority list. On a critical asset with a thin history, that is exactly backwards.
A clean thermal image tells you the asset was not generating abnormal heat on the day, under the load it was carrying. It does not tell you the asset has margin, it does not tell you a connection is not slowly loosening, and it certainly does not tell you it will still be clean under next summer's load. Treating the absence of a documented fault as evidence of low risk is one of the most common reasoning errors I see, and it tends to concentrate exactly on the assets a programme can least afford to get wrong: the essential ones nobody has flagged in years, precisely because nobody has looked hard.
A more defensible basis
A better programme starts from an inventory, not a calendar. Every asset gets an identity, a criticality rating, and a documented test status, so you know not just what was inspected but what was deliberately not, and why. Intervals are then set by consequence and rate of change. The critical, heavily loaded, thinly documented assets are surveyed more often and trended closely. The benign, lightly loaded, well-understood ones are surveyed less often, freeing time and budget for the assets where it matters.
This is the direction modern maintenance standards have moved in general terms, away from fixed time-based routines and toward condition-based and risk-based intervals. The recent shift in electrical maintenance guidance toward condition-based approaches reflects the same logic. Inspect based on what an asset is, how hard it works, and what you already know about it, rather than on the fact that twelve months have passed.
The takeaway
Annual is a fine interval for some assets and a dangerous one for others, and the honest problem is that a single site-wide interval cannot tell the difference. Build the inventory, rate the criticality, factor in the load and the history, and let those set the frequency. And resist the instinct to reward a critical asset for a clean scan by looking at it less. On the assets you cannot afford to lose, a clean image is a reason to keep watching closely, not a reason to look away.
