Evaluation of Parallel Coordinates for Interactive Alarm Filtering

Alarm management is a crucial part of many industrial systems used in generation, transmission and distribution of electric power as well as in production processes, for example in steel mills and oil refineries. Filtering of important alarms from unessential alarms is a critical task for operators of such systems, since failures may result in severe abnormal situations such as plant shutdowns. Currently, alarm filtering tasks are performed with alarm lists that offer limited interactivity. We evaluated an application of Parallel Coordinates for interactive filtering of alarm data by comparing its user performance against typical alarm lists. Statistical results demonstrate marginally significant evidence (p<;0.1) for easy and moderate tasks while there is a highly significant evidence (p<;0.01) for difficult tasks that Parallel Coordinates reduce alarm filtering time. Furthermore, Parallel Coordinates were found to reduce human mistakes.