Thermal imaging produces useful evidence when it's conducted under representative load conditions, documented with sufficient traceability, and positioned as a baseline for future comparison. In commissioning packs reviewed against these criteria, a notable proportion fall short of at least one — and the gap rarely gets flagged because the images are present and the report is signed.
The issue isn't the method. Infrared thermography is a well-established survey tool. The problem is the conditions under which commissioning IR surveys tend to be conducted, and the conclusions that get drawn from them despite those conditions. A clean thermal image means something different depending on the load at the time of capture — and that qualification is frequently absent from the commissioning record.
The load condition problem
Resistive heating at a connection or joint is proportional to the square of the current through it. At 20% of rated load — a common condition during commissioning IR surveys, when not all downstream load is yet connected — a genuinely defective connection may not produce a temperature differential sufficient to be distinguishable from background variation. The image looks clean. The marginal connection enters service without ever having been tested under conditions that would reveal it.
Where load level at time of survey isn't documented on the IR report, the record can't be properly evaluated. A thermal survey at 15% load and one at 80% load produce fundamentally different data sets; treating both as equivalent evidence of connection quality isn't a conservative position. A thermal report that doesn't document operating load at time of capture has limited value for pre-energisation decisions.
Surveys conducted before energisation
A specific and recurring pattern: IR surveys included in commissioning packs that were conducted before the equipment was energised — panels before cables terminated, boards before breakers closed, busbar sections under no load at all. These produce images with no thermal signature beyond ambient temperature. They record that equipment exists and is physically present. They don't record connection quality under any meaningful load condition.
The concern isn't that these surveys were done — there may be legitimate reasons for pre-energisation baseline photography. The concern is when they're included in the commissioning documentation without distinction, making the pack appear more comprehensively surveyed than it is. On an anonymised data centre pre-energisation review, a commissioning pack with fourteen IR survey records across five distribution boards was found to include six surveys conducted before board energisation, covering approximately 40% of the documented coverage. The apparent completeness of the record wasn't reflected in its actual evidential content.
Main LV switchboard survey ≠ downstream distribution coverage. An IR survey of the main incomer and primary switchboard doesn't substitute for coverage at downstream distribution panels and termination points. Each connection at each panel has its own termination quality and load profile. A clean main incomer image says nothing about the downstream distribution.
Point of connection coverage matters. The highest-risk points for resistive heating are terminations and joints, not bus sections. Coverage limited to the main switchboard may miss the distribution points where installation-quality issues tend to concentrate.
Traceability and equipment identification
Thermal images without equipment labels or circuit references are a recurring documentation gap. Where images can't be matched to a specific panel, protective device, or busbar section, they can't be used as a baseline for future comparison surveys, can't support investigation if a fault occurs in service, and can't be cross-referenced against the commissioning test schedule to confirm coverage. The survey may have been conducted correctly — but without identification, it doesn't function as useful documentation.
The baseline purpose of commissioning IR is distinct from ongoing predictive maintenance surveys. A commissioning survey establishes the reference point: this connection, this load, this temperature differential, this ambient. Future surveys identify deviation from that reference. Where the commissioning record doesn't document ambient temperature, load level at time of survey, camera settings, and equipment identification, it doesn't function as a baseline — it records that someone took photographs.
Single-point surveys and their limitations
Single-pass surveys with no intent to repeat have limited evidential value even where they're conducted under appropriate load conditions. The diagnostic power of thermal imaging comes from comparison of the same point over time under comparable conditions. Deviation from baseline is what identifies developing faults. A commissioning survey is the first data point in that sequence. Without a planned follow-up under operational load — typically 12 months post-energisation — the single image tells you less than its presence in the commissioning record implies.
The underlying issue is that commissioning packs frequently treat IR survey records as evidence of installation condition regardless of the conditions under which they were taken. A clean thermal image from a lightly loaded incomer, with no equipment identification, captured before downstream distribution was energised, doesn't demonstrate that the distribution is healthy. It demonstrates that the main incomer was photographed. These are materially different things, and the distinction is worth making before power goes live.