Where exposure monitoring produces numbers instead of protection.
Continuous monitoring supports a hygienist; it can also produce a confident record of the wrong thing. These are the failures we design against, and each has a straightforward remedy.
| What goes wrong | How you notice | The fix |
|---|---|---|
| Instrument placed where it is easy to service | Readings look acceptable while people do not | Measure at the breathing zone of the task. A sensor on a wall near the door describes the door |
| Calibration and bump-testing kept informally | The data is questioned exactly when it matters | Keep the calibration record with the data. An uncalibrated instrument produces numbers, not evidence |
| Alarms with no defined response | People acknowledge the alarm and carry on | Give every alarm level a named action. An alarm nobody is required to act on trains everyone to ignore alarms |
| Averaging that hides the peaks | The shift average is compliant and the exposure is not | Report peak and duration alongside the average — dose is the area under the curve, not the mean |
| Cross-sensitivity ignored | A gas reading tracks something else entirely | Know the interferences of the sensor type before you act on a trend, especially where diesel and blasting products share a sensor |
| Wrong dust fraction | Continuous numbers do not reconcile with gravimetric results | Match the size fraction to the standard your hygienist reports against, or the two records will never agree |
| Ventilation changes not recorded | Exposure trends move for no visible reason | Log fan and ventilation changes with the exposure data; most unexplained trends turn out to be air, not source |
| Area data used as personal exposure | Individual claims rest on a fixed instrument | Keep area and personal monitoring distinct. They answer different questions and are not substitutes |
The two that most often undo a programme are placement and the undefined alarm. Placement decides whether the record describes a person or a room, and it is settled in an afternoon at the start rather than argued about later. The undefined alarm is the slower failure: each unactioned alert teaches the section that alerts are background noise, and that habit outlasts whichever instrument produced it.