In short. Alarms exist to prompt action. Too many alarms, unclear priorities and no procedures mean the important one is missed. Rationalise alarms so each has a cause, a consequence, a priority and a response, and keep an accurate event record for later diagnosis.
What makes a good alarm
- It needs an operator response.
- It is unique: one cause, one alarm.
- It has a defined priority and expected response time.
- It clears when the condition clears.
- It is documented with consequence and action.
Priorities
| Priority | Meaning | Example |
|---|---|---|
| Critical | Immediate action; loss of supply or safety risk | Generator failed to start; UPS on battery |
| High | Action soon; loss of redundancy | One UPS module failed; ATS in manual |
| Medium | Action in shift | Battery charger fault; low fuel warning |
| Low / information | No immediate action | Maintenance due; test complete |
Alarm rationalisation
- List all alarms and their sources.
- Remove duplicates and those that need no action.
- Combine related alarms into one clear message.
- Assign priority, set thresholds and delays to avoid chatter.
- Write a response: what to check, what to do, who to call.
- Review after incidents and at set intervals.
Common alarm problems
- Alarm floods during an event: hundreds of alarms burying the cause. Use first-out indication.
- Standing alarms that are always on and become ignored.
- Chattering alarms that repeat on a noisy signal. Use deadbands and delays.
- Stale alarms for removed equipment.
- Alarms not reaching anyone off-hours.
Event records and sequence of events
An event log with accurate, synchronised timestamps lets engineers reconstruct what happened after an outage: which breaker opened first, which source failed, how long the transfer took. Keep protection relay event records and disturbance records and correlate them. See sequence of events.
Escalation
Define who is notified, how and when: on-site operators, then the duty engineer, then the service provider. Test the notification path, including phones, SMS and email.
Worked illustration
A utility failure triggers 140 alarms in a minute. Operators cannot find the root. After rationalisation, a first-out alarm “Utility failure: generator start commanded” appears at the top, dependent alarms are suppressed while the generator starts, and a separate alarm appears only if start fails. The operator sees one clear message and acts. Same equipment, same event, clearer decisions.
Alarm discipline in critical power
In a loss-of-supply event, an operator sees dozens of alarms within seconds. The system must show what matters first.
| Priority | Meaning | Response time (illustrative) |
|---|---|---|
| Critical | Loss of redundancy or supply at risk | Immediate |
| High | Component failed, redundancy reduced | Within minutes |
| Medium | Degradation or maintenance due | Same shift |
| Low | Information | Review |
Redundancy-loss alarms
The most important alarm in a redundant system is the one that says redundancy is lost. A failed UPS module that is silently covered by the spare leaves the system at N with nobody aware. Alarm on every loss of redundancy and on every standby that becomes unavailable.
Practice
- First-out indication and time-stamped event logs.
- Suppression of consequential alarms during an upset.
- Escalation: if nobody acknowledges within a set time, notify the next person.
- Written response for each critical alarm.
- Monthly review of nuisance alarms.
Out-of-hours coverage
Unattended sites need alarm transmission to a person with authority and the ability to act, plus a documented call-out chain.
See alarm rationalisation and alarms, events and records.
Common mistakes
- Treating every status as an alarm.
- No response procedures.
- Unsynchronised clocks, so event order is unclear.
What to send us
Send current alarm lists and any past incident logs. Send your application details. See also monitoring and SCADA.
Questions engineers ask
How many alarms is too many?
When operators cannot respond to each one. Count alarms per operator per hour in normal and upset conditions and reduce until manageable.
What is a first-out alarm?
It identifies which alarm occurred first in a cascade, pointing to the root cause.
Why do timestamps matter?
Without synchronised time, you cannot tell the order of events across devices.
Should every alarm have a procedure?
Yes for critical and high priority. If no action exists, it may not need to be an alarm.
This guide gives general principles. Apply your site’s operating procedures and standards.