ISA-18.2 alarms: how to eliminate nuisance alarms

Alarms in SCADA should help the operator, not get in the way. In practice, poorly configured systems bury the operator under dozens of activations, among which the truly important one gets lost. The solution is to build the alarm system according to the ISA-18.2 standard. In this article, the engineers at Atlas Scada explain its principles.

The problem: alarm flood

An “alarm flood” is an avalanche of alarms, when during an abnormal situation the system issues so many signals that the operator cannot keep up with sorting through them. Conversely, if alarms fire on every minor deviation and even in normal operation, the operator gets used to ignoring them — and misses a real failure. Both are dangerous.

What ISA-18.2 is

ISA-18.2 is an international standard for alarm management in industrial control systems. It describes how to design, configure and maintain an alarm system so that every alarm is meaningful, clear and requires action from the operator.

Key principles

• Prioritization — every alarm has its own priority (critical, high, warning), so the operator sees at once what to respond to first.
• Acknowledgement — the operator confirms that they have seen the alarm; the system records this.
• Every alarm requires action — if a signal does not call for a response, it is not an alarm but an informational message.
• Logging and analysis — all events are recorded for review and for improving the system.

Alarm rationalization

Rationalization is a review of all alarms: which ones are truly needed, with what setpoints, priorities and delays. Redundant and duplicate signals are removed, and important ones are configured correctly. As a result, the number of alarms drops manyfold, while their value grows.

The alarm lifecycle

The standard considers an alarm at every stage: definition and design, configuration, operation, performance monitoring and continuous improvement. This is not a one-time setup but an ongoing process.

How we do it

In Atlas Scada systems we build the alarm system on ISA-18.2 principles: with priorities, acknowledgement, logging and rationalization. The operator sees what matters rather than a stream of nuisance alarms — and this reduces incidents and human error.

Alarm performance metrics

The standard proposes measuring the “health” of the alarm system: the average number of alarms per operator per hour, the share of time spent in an overload state (alarm flood), and the number of “standing” and frequently recurring alarms. These metrics show whether rationalization is needed.

Relationship to EEMUA 191

ISA-18.2 draws heavily on the EEMUA 191 guide — a recognized industry benchmark for alarm management. Both documents agree on the main point: there should be few alarms, and each should require a clear action from the operator.

Alarm priorities

Priorities are distributed so that critical alarms are a minority and most are warnings. If “everything is critical,” priorities lose their meaning. Proper distribution helps the operator instantly understand what to respond to first.

The key takeaway

A good alarm system is not “as many alarms as possible,” but only meaningful signals that require action. The ISA-18.2 standard helps achieve this through prioritization, acknowledgement and rationalization. This is the approach we apply when developing SCADA and in the Atlas dispatching platform.