Bad value alarms, shutdown-state alarms, several alarms to indicate one event – more information is better, right? Not quite. Before beginning my internship with ProSys, I thought that annunciating more alarms would be better for operations by providing redundancy, clarity, or multiple ways of checking each possible instrument. It was a tempting thought but I soon learned that was not the case. It is more appropriate to say that smaller amounts of carefully chosen information are better than simply more information. For alarms, this is done through alarm rationalization and dynamic alarm management.
It’s useful for understanding alarm rationalization to consider alarms occurring in abnormal, unexpected situations instead of normal situations. For example, a typically hot process stream falling to a low temperature during a regular run state would certainly be unexpected and should receive an appropriate alarm. But why should that same stream give a low temperature alarm during startup if it starts from a cool temperature? According to the CSB, 70% of process incidents occur during startup or shutdown which demonstrates that effective dynamic alarm management is essential for plant safety and performance. Unnecessary alarms can also cause alarm floods and make it difficult or misleading for operators to identify the true problems occurring. Alarm rationalization addresses this issue by eliminating alarms for normal, expected situations and leaving alarms for truly abnormal situations which in turn improves operator knowledgeability and allows plants to meet ISA 18.2 standards.
My internship with ProSys has allowed me to contribute to alarm management projects for clients seeking to implement dynamic alarm management. One major piece of work I did was to create the HMI graphics for operating consoles which display the case logic diagrams that govern which alarms are enabled in the different operating cases. I then linked shapes in the diagrams to data points in the control system such that sufficient instrument values would illuminate the operating case in the diagram, allowing the operator to see the logic-determined operating state and understand which alarms are enabled. The final alarm database and my graphics were successfully implemented near the end of my term.
Towards the end of my internship, I contributed to the front-end design efforts of a new alarm rationalization project. I helped divide a swath of the facility into systems with unique case logic, draft the logic, and select the instruments that would be used to determine cases in the logic. This required an understanding of the process and the equipment for the system breakout as well as the careful inspection of P&IDs to find appropriate instruments.
This experience introduced me to an important niche of process control, provided motivating examples of why alarm management is essential to safe and effective operation, and allowed me to make real contributions to a successfully implemented project. I am grateful to have had this opportunity and for the support from the engineers around me.