Tips for starting an alarm management program

Using the ISA-18.2 standard can help process engineers understand, simplify, and implement a sustainable alarm management program.


Congratulations. You’ve been assigned the task of establishing an alarm management program for your facility. So where and how do you begin? This article presents four practical tips for starting an effective and sustainable alarm management program that conforms to the tenets of a relatively new process industry standard for alarm management published by ISA. 

Tip 1: Understand alarm management terminology, concepts

Plenty of good information on alarm management can be found on the Internet by searching for the terms “alarm management” and “alarm rationalization.” Typically, searching on these terms yields so much information that it is easy to become overwhelmed and confused. Therefore, begin with the basics by becoming familiar with good engineering practices for alarm management (see Figure 1).

Figure 1: A good starting point for building a sustainable and effective alarm management program begins with a clear understanding of objectives and the basic principles of the alarm management lifecycle. Courtesy: Emerson Process Management

Until very recently, process plant owners, process control system manufacturers, and alarm subject matter experts have been practically on their own to coin their own terms and advocate their own alarm management concepts. For example, even basic notions had become highly inconsistent, such as the difference between an alarm and an alert, and the differences among alarm conditioning, suppression, shelving, and inactivation. 

In 2009, ANSI/ISA-18.2-2009: Management of Alarm Systems for the Process Industries was published, also known as the ISA-18.2 standard. The ISA-18.2 standard helps to clear the confusion by providing clear definitions of the common terminology and helping to create a universal alarm management language. It also defines an alarm management lifecycle model, which establishes the recommended workflow processes. This lifecycle provides the central framework for understanding the requirements for building an alarm management program. Any products, services, or in-house processes that are put in place should be mapped against the model to identify gaps and assess compliance. Since the ISA-18.2 standard was published, most alarm solution vendors and industry alarm consultants have begun—or completed work—to align their offerings and solutions with the standard. 

First, understand alarm management terminology and concepts according to the ISA-18.2 standard and then proceed to see what else the Internet has to offer. 

Tip 2: Clarify top-level objectives, program scope

Quite possibly, your new task was motivated by a costly process disruption or maintenance expense caused by a poorly performing alarm system. There are also new and emerging business drivers for creating an alarm management program. ISA-18.2 is the first normative standard suitable for defining good engineering practices for alarm management in the process industries, which should be of interest to regulatory and insurance risk-rating bodies. By applying ISA-18.2 concepts, some companies expect to lower maintenance costs and achieve higher operating performance. For some companies, the business impact of an uncontained abnormal event is so severe that getting and keeping alarms under control is reason enough to create and maintain a standardized alarm management program.

Establishing program goals is important because there should be a natural hierarchy to the scope of a successful alarm management program, which must be matched to the top-level objectives. Proceeding from least to greatest scope, this hierarchy includes:

  • A limited program of nuisance alarm elimination
  • Alarm rationalization with basic alarm redesign as required
  • Advanced alarming techniques and operator HMI optimization.  

Eliminating nuisance alarms: Operators are often subject to nuisance alarms, which are alarms that annunciate excessively or unnecessarily, or do not return to normal after the correct response is taken (e.g., chattering, fleeting, or stale alarms). Prolonged exposure to nuisance alarms can desensitize the operator to alarms and lead them to think it is acceptable to ignore alarms. Also, routine changes in the state of the process (start-up, shutdown, and out-of-service) often can result in alarms that are irrelevant based on the state of the process. However, they still clutter the alarm list. It is not unusual for up to 80% of all alarm activations to originate from a dozen or fewer sources. Thus a program of bad actor alarm identification and elimination can—with minimal effort—significantly quiet the control room, freeing operators from needless interruptions. 

This is just one activity in an effective alarm management program. There is a significant risk to limiting the scope of your program to only nuisance alarm elimination. When a real process disruption or equipment malfunction happens, it often leads to a flood of alarms of varying relevance and usefulness, with alarms prioritized in ways that do not truly reflect their importance. A program of only nuisance alarm rationalization does very little to help the operators cope with actual process disturbances beyond simply lowering the baseline level of background noise. Most alarm management experts caution that, while nuisance alarm elimination is a good thing, it does not equate with a true alarm management program. 

Alarm rationalization, basic alarm techniques: The ISA-18.2 standard specifies that an alarm should be reserved to indicate to the operator that an equipment malfunction, process deviation, or abnormal condition exists that requires a response. The implication is that for every alarm there is a defined operator action that will mitigate or prevent the likely consequences and sufficient time available for the operator to take the prescribed action. 

Alarm rationalization is the process of reviewing potential—and existing—alarms using the guidelines defined in the alarm philosophy (a document you will create) to select alarms for design, and to document the rationale for each alarm. Alarm rationalization thus identifies which alarms to implement, or remove if there is insufficient basis, and their specifications such as priority, limit, and any conditioning such as on/off delays or hysteresis. Clearly, a systematic and careful review of all potential alarm sources, alarm design specification, and subsequent implementation in the control system is a major undertaking. However, this is the necessary investment if one of the top-level objectives is to ensure that the operator can respond decisively and consistently when actual process disruptions or equipment failures occur. 

Advanced alarm techniques, HMI redesign: Advanced alarming is a collection of techniques (e.g., state-based alarming and static/dynamic suppression) that can help manage alarm rates in specific situations. One example is to programmatically consolidate multiple alarm annunciations when a compressor or other large process asset trips offline to prevent the operator from being flooded with alarms and to make clear what has occurred. 

There may also be a need for special control displays to manage such an event. Perhaps a revamp of operator process displays is warranted to increase alarm visibility and promote greater situational awareness. The application of advanced alarming techniques and HMI redesign can represent significant effort and expense. Often, implementing certain advanced alarming techniques can be low-hanging fruit, such as online in-context presentation of alarm guidance based on information collected during the alarm rationalization process. Implementation of advanced alarming and better HMI design can be expensive, so it may makes sense for it to be linked to a DCS upgrade or performance optimization project to deliver on its potential benefit. 

Clarify expectations with your management early, calibrating them to the costs and expected benefits, and scale your alarm management program accordingly. 

<< First < Previous Page 1 Page 2 Next > Last >>

click me