Measurement and traceability9 / 39

Map sensor positions to the load plan and synchronize clocks; do not treat a load-dependent difference as one sensor fault.

Choose a scenario from actual hot, long and frequent-opening operations; do not combine extremes that cannot occur together.

Map the door-area reading by position and time. It is an observation, not product temperature or an automatic disposal decision.

Match unload and next-vehicle load times, then report total, per-run average and longest wait separately.

The calendar alone does not require repeat testing. Compare actual ambient, sun, load, start and route conditions with the validation envelope.

Check product instructions, plate condition and batch records rather than relying on surface frost.

The WHO guide starts from 12 fixed sensors; a medium-sized cold unit typically gets about 20 sensors plus one for ambient.

To separate the container's effect from the vehicle's, read run data crosswise: one container in several vehicles, one vehicle with several containers.

Match device identity to reference, date, result and use decision; link certificate information to the relevant trip record.

Sensor position, response and load produce different measurements; align measurement purpose and method before comparing.

Select sensor points from pilot load geometry and observed hot or cold spots; record location evidence rather than a fixed device count.

Verify timestamps when comparing gaps and excursions near shift handover. Timing can guide review, but does not prove a cause.

Record the container interior, product and coolant separately; do not confuse a starting difference with a route effect.

Trial different fill levels with actual packs; record closure, plate seating, airflow path and handling.

Read month-end excursion counts alongside run volume and data coverage. A fixed week definition makes the comparison more useful.

Match the raw file to logger identity, trip time and time zone; retain checksum, version and data gaps in the decision file.