Back to Blog
Smart Factory

When Every Alarm Rings, No One Listens: Taming Alarm Floods in Manufacturing

When chattering, standing, and duplicate alarms pile up, operators stop listening. This is a practical guide to cutting alarm floods: rationalize alarms against ISA-18.2 and EEMUA 191, then add clear priorities, escalation rules, and alarm data analytics.

POLYGLOTSOFT Tech Team2026-09-157 min read2
Alarm ManagementISA-18.2SCADAEquipment MonitoringSmart Factory

Why Operators Silence the Monitoring System You Just Built

For the first few weeks after a new equipment monitoring system goes live, someone runs to the machine every time an alarm sounds. A month or two later, you find beacons covered with tape and HMI alarm windows minimized. In most cases the problem isn't operator discipline. It's alarm design.

  • Chattering alarms: A signal hovering near its setpoint trips and clears dozens of times a minute
  • Standing alarms: They stay active for more than 24 hours, across shift changes, until nobody pays attention
  • Duplicate alarms: A single drop in air pressure sets off a cascade of ten or more cylinder, sensor, and interlock alarms
  • As this noise piles up, the high-temperature or leak alarm that really matters gets buried. In an alarm flood, usually defined as more than 10 alarms in 10 minutes, operators can't find the root cause. Response slows down, and the result is unplanned downtime.

    The Standards: ISA-18.2 and EEMUA 191

    ISA-18.2 (IEC 62682) treats alarms as something you manage over a lifecycle, not as setpoints you configure once. The cycle runs from alarm philosophy → identification → rationalization → detailed design and implementation → operation and maintenance → monitoring and assessment → management of change → audit.

    These performance benchmarks are widely used:

  • About 150 alarms per operator per day (6 per hour) is manageable. More than 300 per day means the operator is overloaded
  • The plant should spend less than 1% of operating time in flood conditions
  • Aim for a priority split of roughly 80% medium / 15% high / 5% critical
  • EEMUA 191 likewise treats about one alarm every 10 minutes as a desirable average rate. Factories often log more than 1,000 alarms a day, which puts them far from these targets.

    Alarm Rationalization in Practice

    Rationalization is done in workshops with production, maintenance, and quality staff in the room. For every alarm, the team defines four things:

  • Cause: What triggers it?
  • Consequence: What happens if it's ignored?
  • Operator action: What exactly should the operator do?
  • Allowable response time: How many minutes are there to act before the consequence occurs?
  • If nobody can write down an operator action, the alarm is downgraded to an event log entry. Chattering is handled with a deadband (for example, 2–5% of the setpoint) and on/off delays (for example, 5–15 seconds). State-based suppression hides related alarms while equipment is stopped or being set up, which removes the duplicates.

    Designing Priorities and Notification Paths

    Priority is set by how severe the consequence is and how much time there is to respond:

  • Critical: Safety risk or major equipment damage → beacon and horn on the floor, plus mobile alerts to the maintenance lead and shift supervisor at the same time
  • High: Risk of quality defects or a line stop → HMI pop-up plus a mobile alert to the assigned operator
  • Medium: Preventive action needed → shown in the HMI list and included in the shift handover report
  • Set escalation rules in advance too. For example, if a critical alarm goes unacknowledged for 5 minutes, the supervisor is notified automatically. After 15 minutes, the plant manager is notified.

    Turning Alarm Data into an Improvement Asset

    Alarm history is the cheapest data you have on equipment health.

  • Top 10 frequent alarms report: The top 10 alarms often make up more than half of all alarms. Fixing just those 10 can cut the alarm load sharply
  • Alarm-to-downtime correlation: Look for alarm patterns that repeat in the 30 minutes before a stoppage and use them as leading indicators
  • Link to improvement work: Send recurring alarms on the same asset to maintenance as work orders, and send quality-related alarms to process engineering as improvement tasks
  • A Phased Adoption Roadmap

  • Baseline (2–4 weeks): Measure alarms per operator per hour, time spent in floods, and the number of standing alarms
  • Clean up the worst offenders: Rationalize the top 10 first to show results quickly
  • Document the alarm philosophy: Write down priority criteria, suppression rules, and the change approval process
  • Audit regularly: Review KPIs every quarter and re-examine new or modified alarms
  • POLYGLOTSOFT builds alarm management without replacing your SCADA or MES. We add an alarm analytics layer on top of the systems you already run. It can be rolled out in stages: alarm history collection, Top 10 and flood dashboards, role-based mobile escalation, and integration with maintenance work orders. If alarm noise is wearing down your floor, get in touch and we'll start with an assessment of where you stand today.

    Need Technical Consultation?

    Our expert consultants in smart factory, AI, and logistics automation will analyze your requirements.

    Request Free Consultation