Avoiding Abnormal Situations with Better Alarm Management

Earlier this month my DeltaV News RSS feed announced the availability of an alarm management whitepaper written by the ARC Advisory Group. The paper, Emerson Strategies for Abnormal Situation Avoidance & Alarm Management, describes an issue many process manufacturers face in dealing with too many alarms. This makes it hard for their operators to distinguish between critical impending abnormal events and nuisance alarms.

The Emerson approach to address better management of alarms is based upon ARC’s Six-Sigma DMAIC (Define, Measure, Analyze, Improve, Control) model and

…includes data collection, statistical analysis, alarm evaluation, and system improvement – all within the context of ongoing evaluation and continuous improvement.

Relative to alarm management, the whitepaper describes the DMAIC process:

Define relates to philosophy development, Measure relates to determining alarm behavior and alarm effectiveness, Analyze relates to root cause analysis and performance benchmarking, and Improve relates to the remedial action necessary to align the prevailing implementation with the alarm philosophy. Finally, Control relates to alarm execution.

Advancing technology is a source of the alarm proliferation. The authors note:

In the days of hardwired controls and alarms, engineers were very stingy with alarms, in part because each alarm point had a cost. The primary issue with alarm systems is there is too much information for an operator to assimilate and act on. Ten years ago, it cost about one thousand dollars to add an alarm. Current automation systems have essentially eliminated the cost of adding more alarms and, therefore, the incentive to limit or rationalize their number.

I have an upcoming post that discusses the importance of up-front planning your alarm strategy when you’re planning your project’s functional requirements. But what do you do in an existing facility with an alarm overabundance?

The whitepaper addresses building a business case to justify and alarm management project. The keys areas build the business case include safety, unplanned downtime, better information management and reduced troubleshooting time, and changing the role of the operator toward higher-value activities.

ARC notes that unplanned shutdown costs process manufacturers on aggregate between 2 percent and 5 percent annually. This may be a justification opportunity by looking at root causes of unplanned shutdowns in your plant. A review of the alarm and event logs around these incidents can reveal the number alarms the operators saw and the actions they took as a result.

The whitepaper also addresses the important role of the EEMUA Publication 191–Management of Process Alarms in developing your alarm management strategy. A guiding principle described in EEMUA 191:

…a usable alarm system must be relevant to the user’s role at the time, indicate clearly what response is required, and be presented at a rate the user can deal with, and be easy to understand.

With this backdrop, the paper explores the applications within Emerson’s DeltaV system like the Event Chronicle, integration with 3rd-party alarm management applications via the OPC alarms and events communication standard, and the DeltaV Analyze alarm analysis program. Coupling these applications with alarm management services can help process manufacturers through the process of data collection, statistical analysis, alarm evaluation, and system improvement.

And much like the safety lifecycle as defined by the IEC 61511 international safety standard, ongoing evaluation of the overall alarm strategy is important throughout the lifecycle of the plant.

Posted Wednesday, February 27th, 2008 under Alarm Management, Operator Performance.

2 comments

  1. Javier Arellano says:

    Here we have a key factor that you quote “The primary issue with alarm systems is there is too much information for an operator to assimilate and act on.” Why is so important the alarm management? To me, it is beacuse if you put 1000,000 alarms in a system, the operator will be overwhelmed with info and it is possible that can not react properly if a situation takes place. Kindly Regards.

Leave a Reply