By Doug Metzger
Since the early to mid '90s there has been an increasing recognition within the process industries that alarm systems, particularly in distributed control systems (DCSs), have gotten out of hand. Results of studies by the ASM Consortium (www.asmconsortium.com) documented in 1995 (Nimmo, 1995) concluded at that time that better handling of abnormal situations in the U.S. petrochemical industry alone could save up to $10 billion per year―a figure that has been repeatedly quoted in literature. Alarm systems have often been identified as part of the problem, rather than the solution. Because of the ease of adding alarms in a DCS, and the apparent added security of doing so, the numbers of alarms an operator must deal with have grown over the years from a few hundred to many hundreds or thousands (Andow, 2000). This often presents the operator with excessive, redundant and unnecessary alarms, overloading the operator and obscuring real problems. As Bransby observed, during plant upsets "there are real risks of important alarms being missed by the operator―with potentially severe consequences." (Bransby & Jenkinson, 1998, p. 62). Unfortunately, experience has shown that there is not a single quick and easy solution to the problem.
Work by the Engineering Equipment and Materials Users Association (EEMUA, www.eemua.org) in the 1990s resulted in publishing its guidelines in 1999―EEMUA Publication No. 191, Alarms Systems, a Guide to Design, Management and Procurement. This document and its update in 2007 (EEMUA, 2007) have been the premier guide for the emerging worldwide consensus on principles and practices for alarm management in industrial process control. Along with documenting the basic principles of good alarm system management and successful techniques in use, EEMUA Publication No. 191 set some clear alarm system performance targets―among them, less than one alarm per 10 minutes as "very likely to be acceptable" and less than ten alarms in 10 minutes following an upset as "should be manageable, but may be difficult" for the operator.
EEMUA performance targets, though aggressive, based on study of industry norms (Bransby & Jenkinson, 1998), and the core principles of alarm management established in Publication 191, have had a normalizing influence on the field of alarm management in the process industries.
The ASM Consortium
The Abnormal Situation Management Consortium (ASM Consortium) began in the early1990s, as an outgrowth of issues around alarm management in distributed control systems. The Consortium's focus of "abnormal situations" is broader than just alarms, but alarm management remains a central theme.
As part of its activities, the ASM Consortium had supported EEMUA's efforts to produce and publish its guidelines in 1999 (and later in 2007). Realizing early in this decade that the EEMUA performance targets were quite aggressive, the Consortium conducted human factors research which showed that EEMUA targets were, in fact, appropriate from an operator loading standpoint (Reising, Downs, & Bayn, 2004). Meanwhile, a study of Consortium member sites (Reising & Montgomery, 2005) showed that (a) such averages can be achieved, but that (b) most operating plants were not meeting these numbers, some off by an order of magnitude or more.
The EEMUA Publication 191 provided a very good basis for alarm management practices; but in a desire to achieve the EEMUA targets, the Consortium members felt the need for more specific recommendations and examples drawn from the collective experiences at member company sites. Like EEMUA, Consortium members wanted to produce guidelines and not standards. The ASM Consortium guidelines on effective alarm management practices were first created in 2003 for use within the Consortium membership, and have undergone six revisions since the first release. The current release, completed in 2009 (Errington, Reising, & Burns, 2009), was made available for purchase by the public on June 15.
The ISA18 Committee
Unlike EEMUA and the ASM Consortium, production of standards documents is one of the missions of the International Society of Automation (ISA, www.isa.org). In June of 2003, the ISA Standards and Practices Board reactivated its SP18 Committee (which had produced a standard in 1979 on Annunciator Sequences and Specifications―ISA-18.1-1979) to create standards for computer-based alarm systems for the process industries.
The first full committee meeting was held in June of 2004 (a skeleton group had met in October of 2003), and from then until now has been comprised of a cross section of experienced personnel from process control and alarm systems producers, users, consultants and engineering contractors. There are currently around 90 members on the committee, 25 of whom are voting members. Members meet face-to-face twice yearly and participate in the interim by editing and commenting on the drafts, both formally and informally. According to Nicholas Sands, co-chairman of the committee, there have been eight interim drafts, and since May of 2006, over 8,000 formal comments have been received and tracked to resolution by the committee. It was approved by committee vote in April 2009, the ANSI/ISA approval process was completed on June 23, and it was made available for purchase in July of this year (ANSI/ISA-18.00.02-2009).
The ISA18 committee was intentionally comprised of known experts in the field, including some with ties to the earlier work by the ASM Consortium and EEMUA, so that the results would be consistent with and a part of the emerging worldwide consensus on alarm management principles and practices.
Different Approaches, Same Target Problem ASM Guideline Document Structure
The structure and content of the ASM guideline document was driven by a number of factors: