Lifecycle Alarm Management for Performance Improvement

Understanding the problem of alarm management must begin with a sense of the alarm’s sound.

Aw 1757 1006 Team
In the control room of a modern chemical plant running under normal conditions, the alarm horn will sound once or twice in every 10-minute period. Under upset conditions, engineering guidance considers one or two alarms every minute to be manageable; but many alarm systems cannot meet this standard and will generate far more. During a true emergency, thousands of alarms will sound, coming so fast that no meaningful response to all of them is possible.

The consequences are significant. Financial losses associated with poor alarm systems amount to tens of billions of dollars annually. Operators cannot optimize complex production processes when they are continually responding to alarms, many of which are irrelevant. Buried in noise and misdirection, critical alarms can go unnoticed. Time spent operating under upset conditions lengthens. The frequency of plant shutdowns forced by safety instrumented systems increases. In the extreme, the result is catastrophic failure.

So why not fix it?  In many plants, alarm management problems have been fixed—and fixed again—only to return as bad or worse than before. It is a pattern seen often enough that industry experts conclude that, although the tactics are often right, the strategy is wrong. “Alarm management problems will not be solved by alarm improvement projects alone,” says Nicholas P. Sands, CAP, PE. Sands is co‑chair of the International Society of Automation’s (www.isa.org) committee on ISA‑18.2 (2009), one of industry’s principal standards covering the design and management of alarm systems.

Continuity Lacking

Essentially, discrete solutions are ill‑equipped for solving continuous problems. Projects start and stop; project teams come and go. Without some organizing structure to link them together, however well they are conceived and executed, independent projects lack continuity. Between improvement projects, the alarm system inevitably degrades. “Only a true lifecycle approach to alarm management will retain the benefits of alarm improvement projects. This approach becomes as much a part of the alarm system as any of its physical components,” says Sands.

A lifecycle approach recognizes that alarm system performance depends on the definition of solid system requirements (referred to as an “alarm philosophy”) and on the cumulative effect of a series of actions repeated cyclically over time. The lifecycle encompasses design, training, operation, maintenance, monitoring and change management. According to Kim Van Camp, ISA‑18.2 committee representative for controls vendor Emerson Process Management (www.emersonprocess.com), in Austin, Texas, the best way to understand why the lifecycle approach works is to conceptualize it as a classic, iterative, feedback control loop. The setpoint is the alarm philosophy, which is the requirement that the alarm system must satisfy where it is (or will be) deployed. Feedback comes from ongoing monitoring of system performance relative to the alarm philosophy. Lastly, the people who design and operate the system are the mechanism of the loop’s control action.

“A true lifecycle approach requires us to think more broadly than we often do,” says Peter Andow, a former professor of chemical engineering who is now principal consultant for advanced solutions in Europe, the Middle East, and Africa for Honeywell Process Solutions (www.honeywell.com), the Phoenix-based automation vendor.  For example, where training is concerned, develop curriculums for design engineers and operators that would graduate specialist operations engineers, because running a distillation column is different than designing one.

“Considering the entire lifecycle also reveals the interfaces between our systems and their surroundings,” says Andow. “It allows us to see alarm systems in the context of fundamental process design, and to see the plant in the context of the community around it—to take account of the interactions.”

Marty Weil, martyweil@charter.net, is an Automation World Contributing Writer.

International Society of Automation
www.isa.org

Emerson Process Management
www.emersonprocess.com

Honeywell Process Solutions
www.honeywell.com

Subscribe to Automation World's RSS Feeds for Columns & Departments

More in Control