The alarm system is a daily encounter for most operations across different industries. For oil and gas, the alarm system alerts and reminds the operation to take action(s) to revert the process back to the normal operating range. For pharmaceuticals and batch operations, the alarm system can be used to inform the operator to shift to different phases in the production. For hospitals, the alarm systems can help the nurses and doctors when patients are experiencing difficulties and need attention. 

The ISA 18.2/IEC 62682 contains clear guidelines on how to define the alarm system; the conducts and the deliverables of rationalization workshops; and the audit and monitoring of the alarm system during operation. However, each of the steps, which lead to each phase in Alarm Management, starts with data.

Pre-rationalization data management

Before we start the alarm rationalization workshop, we need to retrieve data from the control system, the P&IDs, and various documents. These data need to be in a format that can be compatible with the import process back into the control system after the rationalization. The preservation process of such data format is absent from the standard. We often need to borrow concepts from Data Science for guidance on the upkeep.

Master Alarm Database

The Master Alarm Database (MADB) is mandated as the deliverable of the Alarm Rationalization workshop. It contains all the vital and authorized settings of the alarm system. 

However, the standard does not provide guidance on the data structure, data integrity, and the upkeep of the database. These issues can be addressed from the disciplines of Data Science. 

Alarm System Performance Monitoring and Assessment

After the alarm settings have been uploaded for operation, the standard suggests monitoring and assessment of the alarm system for performance issues. This process involves digging through various reports to look for Key Performance Indices (KPIs) and identify different alarm problems, such as Bad Actors, Stale Alarms, or Chattering Alarms. 

Interpreting data and deriving solutions are the main goals for Data Scientists/Data Analysts. We can certainly borrow some of the ideas from the discipline of Data Science.

The relationship between Alarm management and Data Science is far more complicated than it appears. I attempt to investigate a bit more on this topic in the webinar below.

As always, Stay safe.