MAR-APR 2018

Issue link:

Contents of this Issue


Page 19 of 57

20 INTECH MARCH/APRIL 2018 WWW.ISA.ORG PROCESS AUTOMATION procedure. A managed H2S class alarm will detect that change during an audit, recognize the sampling procedure is sues, and result in a safer condition. Steps for an effective alarm system Falling back on the old adage that you cannot improve what you do not mea- sure, monitoring, assessment, and audit activities are arguably the most impor- tant for long-term success. By undertak- ing monitoring and assessment, actual performance is measured and available for improvement. Similarly, audits of the work processes highlight any behavior that does not follow the alarm philosophy. Note: The ISA technical report on this topic (ISA-TR18.2.5-2012, Alarm System Monitoring, Assessment, and Auditing) is a good reference for methods, metrics, and work practices. The three loops shown in the ISA life cycle (figure 1) highlight areas for focused activity: audit and philosophy, monitor- ing and management of change, and monitoring and maintenance. l An audit is conducted to ensure that the alarm management work pro - cesses are sound and aligned with the philosophy. As alarm manage - ment work processes mature, it is important to make sure that the philosophy is updated to reflect changes in practice. Philosophy doc - uments that remain out of sync will fall increasingly out of use. l In monitoring performance, issues related to following management of change procedures will become ap- parent. To effectively manage alarm systems, it is common to develop "fast track" or simplified processes. Allowing changes to occur without following a formal management of change process will make it difficult to achieve system integrity. l Similarly, allowing the maintenance function to transfer alarms in and out of service without structured expectations and schedules can cause less effective systems. Manage the work processes Alarm class can help manage each stage of the life cycle: Philosophy: In the philosophy process, operational definitions or terms and work processes are set. Alarm classes are listed and defined. This establishes clear ex- pectations about alarm classes and how their related requirements are managed throughout the life cycle. Identification: In a similar manner, clear alarm class ground rules set in the work practices related to identification help manage consistency. Rationalization: Guidelines relat ed to different classes of alarms help stream- line and manage the rationalization process. Classes are assigned during rationalization. Detailed design: Alarm classes may have specific requirements for setting the alarm design basis. This may be related to things such as the alarm limit or priority, implementation details, the general pre- sentation on the HMI, or the need for spe- cific online help information. Implementation: When implement- ing certain alarm classes, specific re- quirements for testing and training may be required. Operation: When certain alarm classes are in operation, they require refresher training. Maintenance: Taking certain alarm classes out of operation and placing them into the maintenance stage of the life cycle may require specific remediation plans. Additional monitoring require - ments, altered modes of operation, or specific testing requirements might be needed before returning to service. Monitoring and assessment: An emerg- ing concept requires the alarm system as a whole to meet certain performance levels to continue to take credit for independent protection layer (IPL) alarms. Example alarm class: IPL alarms A layer of protec- tion analysis may identify alarms that provide risk reduction. These are generally known as independent protection layer alarms. Such alarms may have additional requirements for monitoring, including frequency, time in alarm, time shelved, time out of ser- vice, average alarm rate when active, and percent of time participating in an alarm flood. "High frequency" may indicate time durations higher than estimated during the functional safety studies. "Time in alarm" may reflect exposure time to the underlying hazard. "Shelved" and "out of service time" may reflect time not avail- able as a layer of protection. Review of these time frames in an audit may re- flect the need for recognition of different modes of operation, addition of alternate alarms, and related recordkeeping Management of change: Alarms gen- erated from engineering studies or func- tional safety may require different levels of approval for changes. Audit: Certain classes of alarms may require specific audit periods or level of detail on audits to meet compliance requirements. Right size the effort Adding alarm classes may seem daunt- ing at first. However, managing the re- quirements is much easier to implement if you recognize that some alarm types have different life-cycle requirements. Segregate those alarms into classes as a first phase. To "right size" the effort, use an initial approach of generating a set of alarm classes that splits out the key "special" or HMA alarms and leaves the rest as a general class. It is important to approach the process by finding the important alarms with special require - ments, rather than considering the need to discuss and classify every alarm. n ABOUT THE AUTHOR Bridget Fitzpatrick (Bridget.Fitzpatrick@, an ISA Fellow, is disciple tech- nical authority for HMI, abnormal conditi on management and human factors, for the automation and control organization within Wood. Wood provides project, engineering, and technical services for energy and indus- trial markets. Fitzpatrick has an MBA in tech- nology management from the University of Phoenix and a SB in chemical engineering from the Massachusetts Institute of Technol- ogy. She sits on the ISA Standards and Prac- tices Board and is managing director of the ISA18 (Alarm Management) committee. View the online version at

Articles in this issue

Links on this page

Archives of this issue

view archives of InTech - MAR-APR 2018