Alarm Flood Analysis (EEMUA 191 / ISA-18.2)

Description

Measures alarm and notification rates against benchmarks, identifies floods, chattering and stale alarms, and rationalises each alarm against a required operator response.

Process industries learned that operators drown when alarms exceed what a human can process. EEMUA 191 set benchmarks: an average of about one alarm per 10 minutes per operator is acceptable, more than 10 alarms in 10 minutes constitutes a flood, and every alarm must require a defined operator action. Analysis uses alarm logs to compute rates, top-ten 'bad actors', chattering alarms (repeated within seconds), standing alarms and priority distributions (recommended roughly 80% low, 15% medium, 5% high). Rationalisation then reviews each alarm's cause, consequence, required response and time available. Notifications, badges and AI alerts in digital products suffer exactly the same pathologies, and this is a ready-made quantitative method.

Attributes

Loss type
Unmigrated
Effort
Days

Steps

  1. Export notification or alert logs per user or operator with timestamps, type and priority.
  2. Compute average rates per 10 minutes and per hour, peak 10-minute counts and the percentage of time in flood (>10 per 10 minutes).
  3. Identify bad actors: the top 10 sources by volume, chattering alerts (repeats within 60 seconds) and standing alerts never cleared.
  4. Check priority distribution against roughly 80/15/5 low/medium/high and look for priority inflation.
  5. Rationalise each alert in a workshop: what action is required, how quickly, what happens if ignored; delete or demote alerts with no required action.
  6. Re-measure rates and user response times after changes, and set ongoing monitoring targets.

Sources

  • EEMUA Publication 191, Alarm Systems: A Guide to Design, Management and Procurement (3rd ed., 2013)
  • ANSI/ISA-18.2, Management of Alarm Systems for the Process Industries (2009/2016)

Also known as

  • Alarm rationalisation
  • Alarm management
  • Notification load analysis

Categories

  • Accessibility
  • Human Factors & Error