An alarm is not a control system. It is one link in a control system. Incident investigations repeatedly show that alarms were generated, but did not lead to a timely response: contact details were outdated, the alarm had been switched off, or nobody felt responsible.

The two ways alarm management fails

  • Too many alarms. Door openings, defrost cycles and badly set limits create noise. People start ignoring messages, and a real event gets lost.
  • Too few people reached. A message goes to one phone that is off, and nothing else happens.

Design principles

  1. Base limits on the material. Start from what the product or sample can tolerate, then set warning and alarm levels.
  2. Use delays deliberately. A delay filters expected disturbances, but it must be shorter than the time the material can safely be exposed.
  3. Escalate until acknowledged. Define a sequence of contacts and channels. Escalation should continue until a named person accepts the alarm.
  4. Record the response. Who acknowledged, when, and what was done belongs in the record.
  5. Test regularly. Test the full chain, including phone numbers and rosters, not just the sensor.
  6. Monitor the monitor. Loss of a sensor, a monitoring station or connectivity must itself raise a technical alarm.

Review and improve

Alarm statistics show where limits are too tight, which assets are unstable and how long acknowledgement takes. A periodic review turns alarm data into fewer, better alarms.

Who does what when an alarm fires

XiltriX Intelligence applies the limits, delays and escalation sequence agreed with the customer, and keeps escalating through the customer's contacts until someone acknowledges. The customer's own team assesses the situation and carries out the physical response.
XiltriX does not respond to the customer's alarms. XiltriX Care provides 24/7 technical support when the monitoring system itself has a problem.

How XiltriX Resilience fits

XiltriX Resilience is an Operational Resilience System built on three layers:
  • Infrastructure: independent sensors and monitoring stations with battery support, resilient connectivity and local data buffering, so a power or network interruption does not create a gap in the record.
  • Intelligence: validated software that turns measurements into alarms, escalation, trends, reports and audit-ready evidence.
  • Care: installation, validation, calibration, preventive maintenance and 24/7 technical support for the monitoring system, delivered by one partner that is ISO 9001 and ISO 27001 certified.

Where to go next