Sensor placement principles for mapping studies

Placement is the design decision that determines what the study can and cannot show. Three questions guide it.

Where is product?

Sensors go at product height, in the air product sees. A sensor taped to a steel rack or placed on the floor measures the rack or the floor.

Where is risk?

• Doors, loading bays and hatches.
• Evaporators, heaters, air-handling outlets and returns.
• External walls, windows, roof, ground floor.
• Top of racking (warm air) and bottom near the floor (cold air).
• Areas with poor circulation behind columns or dense storage.

Where will monitoring live?

Include the existing or intended monitoring positions so the mapping data can show how well they represent the worst-case locations. Give every sensor a unique identifier, a described location and, if possible, coordinates — the analysis needs them, and a spatial map cannot be produced without them.

A worked plan for a 2–8 °C cold room

The three questions above produce a specific list once applied to a real space. For a walk-in cold room of roughly 4 × 3 m with racking on two walls and a single door, a defensible plan runs to twelve to fifteen sensors and looks like this.

Note the last two. The sensor beside the monitoring probe is what lets the study answer the question the monitoring system exists for — does the permanent probe see what product sees? Without it, the mapping and the monitoring are two unrelated datasets. The ambient sensor costs one logger and explains most of the surprises in the data.

Positions that look sensible and are not

• Taped to a steel upright or a rack beam. Metal has high conductivity and a thermal mass of its own; the sensor reads the rack, lagging and smoothing the air it was supposed to measure.
• On the floor or against a wall. Both are surfaces with their own temperature, and neither is where product sits.
• Inside a pallet or buried in a carton, unless the protocol is explicitly characterising product temperature rather than air.
• In the direct discharge of an air handler when the intent was to measure the room. That position is legitimate as a deliberate extreme, but it must be labelled as one and usually excluded from acceptance.
• Near a light fitting that is on during the study and off in service, or vice versa. Record the lighting state either way.

Metadata is part of placement

A position that is not recorded does not exist for the analysis. Every sensor needs a unique identifier that matches the one in the data file, a location description a stranger could follow, and where possible coordinates — x, y and height. Photographs of each position, taken at placement, settle more arguments than any other single piece of evidence, and they take ten minutes.

Write the rationale for each position into the protocol as you plan it. The list of positions answers what; only the rationale answers why, and why is what an assessor asks.