I have been following the subject of alarm management in a DCS (Distributed control system) with great interest. In many process incidents that I investigate (batch and continuous processes), I observe that the flood of alarms that appeared during the emergency effectively negates the usefulness of the alarm. The irony is that we bring in the DCS with its enormous capability and then realize that alarm management in the DCS is a big issue. The root cause of the problem is the misuse of the enormous capability of the DCS. When installing the DCS it is human tendency to assign all probable alarms, thinking that we are using the capability of the DCS! Now the International Society of Automation has brought out a standard called ISA 18.2 - management of alarms in process industries on June 23rd,2009. The definition of alarm as stated in the standard is "an audible and/or visible means of indicating to the operator and equipment malfunction,process deviation or abnormal condition requiring a response". How did we manage to forget this definition?
I was part of a team of process engineers and process operators in a World Scale methanol plant, where we sat down identifying whether each alarm that was provided was really necessary or not. To our surprise, at the end of the exercise, we had reduced the alarms by over 50%!. Today, there are alarm suppression software sold by various vendors, but I feel the solution is simple - DO NOT COMPLICATE THINGS AT THE BEGINNING ITSELF BY PROVIDING TOO MANY ALARMS!
To achieve the goal process Safety I am of the opinion that things must be kept simple. In other words, KEEP IT SIMPLE & SAFE (KISS!)
While on the topic of alarm management, I have also observed the other extreme in few cases - there weren't enough alarms provided. Such issues crop up not in process plants but in storage and transfer facilities where enough thought was not applied in identifying the alarms required.
No comments:
Post a Comment