Alarm Overload…Nuisance Alarms…Alarm Floods…Incorrectly Prioritized Alarms…. These alarm management problems are all too common in the modern Distributed Control System (DCS).

Why is this?  In the “olden” days (read panel boards and alarm lightboxes), there was considerable thought put into what alarms were necessary because there was limited real estate available and an actual cost to implementing them (approx. $1000 per alarm). Fast forward to today when alarms are “free.” The modern DCS provides alarms galore. A typical analog indication block provides:
• High
• High-high
• Low
• Low-low
• Rate-of-change
• PVBad
• More alarms out-of-the box

It’s simple to enable all alarms available without considering which are really required. When it comes to alarms, more is better, right?...

How do we fix it?  We need to return to the days of careful consideration of whether an alarm is truly necessary. This is a process called alarm rationalization, which is a proven alarm management technique and one of the key stages of the alarm management lifecycle defined in the ISA-18.2 standard.  As an alarm will result in a direct consequence if a defined operator action is not taken, rationalization forces us to consider, and document the consequence and corrective action (among other things). If this cannot be defined, then the alarm is no longer valid or necessary.

Try your hand at alarm rationalization. Check out the example in the March issue of Control Engineering: “Manage Alarms Using Rationalization.”