EEMUA Publication 191 is the document that first put hard numbers on what a good alarm system should look like, and it remains the reference most control engineers reach for when they want to judge whether an operator is being helped or drowned. Published by the Engineering Equipment and Materials Users Association, a UK-based industry body, it is written as guidance rather than as a formal certifiable standard. This page explains what EEMUA 191 actually contains, the benchmark alarm rates it made famous, and how it sits alongside the later ISA-18.2 standard.
EEMUA 191 in one line: EEMUA 191 is a UK industry guidance document, formally titled Alarm Systems: A Guide to Design, Management and Procurement, that defines what an effective alarm system should achieve and offers benchmark performance metrics for judging one. Its best-known figures are target alarm rates - on the order of one alarm every ten minutes for an operator in steady-state operation - and it is a recommendation set rather than a mandatory standard, which distinguishes it from ANSI/ISA-18.2.
EEMUA 191 grew out of the recognition, after several high-profile process incidents, that operators were being presented with far more alarms than any human could realistically act on. Rather than prescribe a rigid procedure, the publication describes the human factors behind alarm response, sets out design principles for how alarms should be selected and presented, and gives management practices for keeping a system healthy over its life. It treats the alarm system as something to be designed deliberately and then measured, not something that simply accumulates as instruments and control loops are added.
A large part of the document's value is that it defines what an alarm is for in the first place: an alarm should indicate a condition that requires a timely operator response, and every alarm should have a defined action the operator is expected to take. Signals that carry no required action - status indications, log-only events, or nuisance repeats - are, by this definition, not alarms and should not be annunciated as such. That single principle underpins most modern alarm rationalization work, where each candidate alarm is tested against the question of whether an operator can and must do something about it.
EEMUA 191's most cited contribution is a set of target metrics that turned a fuzzy sense of overload into measurable goals. The headline figure is the steady-state alarm rate: a well-performing system should present an operator with roughly one alarm every ten minutes on average during normal, stable operation, which the guidance frames as a manageable and acceptable long-term load. It also puts bounds on peak load during upsets, describing how many alarms in the first ten minutes after a disturbance are considered acceptable versus overloading, since it is exactly during a plant upset that a flood of alarms does the most harm.
Alongside those rate figures, the publication popularized other now-standard health indicators: the proportion of time an operator spends in alarm flood, the number of standing or long-standing alarms sitting on the display, the share of alarms attributable to a handful of bad-actor points, and the distribution of alarm priorities. These metrics are deliberately expressed as benchmarks to compare a real system against, not as pass or fail thresholds, because the guidance recognizes that context and plant type vary. Their lasting influence is that almost every alarm performance report produced today, and every alarm analytics package, traces its key numbers back to the categories EEMUA 191 laid out.
The metrics EEMUA 191 defined map directly onto what a modern SCADA or cloud monitoring platform records automatically. Because every alarm event carries a timestamp, tag, and priority, the system already holds everything needed to compute average alarm rate, count standing alarms, rank the worst-offending points, and flag periods that exceed the flood threshold. A platform such as Merobix, which continuously logs alarm activity across remote oil and gas sites, can present these EEMUA-style figures as a running scorecard rather than something an engineer has to reconstruct by hand from historian exports.
This matters most on unmanned and lightly staffed sites, where the person responsible may be watching many locations from a single dashboard and cannot afford to have real problems buried under repeating chatter. Applying the EEMUA 191 benchmark of one alarm per ten minutes as a design goal keeps the remote operator's attention meaningful: when an alarm arrives, it is likely to be worth acting on. Sites that drift well above that rate are, by the guidance's own logic, no longer alerting effectively.
It is worth stressing that EEMUA 191 is guidance, so a SCADA system does not become compliant with it in a certifiable sense; rather, it uses the publication's benchmarks as targets and its principles as a design discipline. In practice, teams often adopt the EEMUA 191 metrics as their key performance indicators and then use the more formal ISA-18.2 lifecycle to structure the actual work of getting there.
EEMUA 191 is a guidance document, not a formal certifiable standard. It sets out recommended practices and benchmark performance figures rather than mandatory requirements, so a plant is measured against its targets instead of being certified compliant. It is often used together with ANSI/ISA-18.2, which provides the formal standard framework.
Its best-known benchmark is that a well-performing alarm system should average about one alarm every ten minutes for an operator during steady-state operation. It also describes acceptable versus overloading alarm counts in the first ten minutes after a plant upset, since peak load during a disturbance is when excessive alarms are most dangerous. These are benchmarks for comparison, not rigid pass-fail limits.
EEMUA 191 is UK industry guidance focused heavily on human factors and benchmark metrics for judging alarm system performance. ISA-18.2 is a formal North American standard that defines a structured alarm management lifecycle from philosophy through audit. The two are complementary: many teams adopt EEMUA 191's metrics as targets and use the ISA-18.2 lifecycle to organize the work of achieving them.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.