Process Event Avoidance: Using Identification of Process Faults to Avoid Abnormal Process Events
- Type: Conference Presentation
- Conference Type:
AIChE Spring Meeting and Global Congress on Process Safety
- Presentation Date:
March 29, 2017
The Center for Operator Performance commissioned the University of Texas at Austin to examine new ways to identify process faults and predict events. This research developed a geometric analysis tool that beat the broadly used benchmark Tennessee Eastman Challenge Process published average detection time for twenty faults by over 90%. The geometric analysis approach detected three faults that were either missed or signaled far too late for corrective action by 12 other published techniques.
The geometric tool developed by the Center for Operator Performance and the University of Texas at Austin has shown the capability to identify process faults as leading indicators of abnormal process events with a very low rate of false positives. The technique has been successfully applied to detection of surge in a large, multistage compressor, where analysis of past surges helped identify event signatures as early as a few hours before surge onset. It has given equally impressive fault detection for distillation flooding events and for detecting precursors to flaring events, continuing to show a low false positive rate.
In addition to developing the analysis tool, the research project has provided visualization techniques for process monitoring in operator displays, abnormal event prediction, and identification of process variable(s) causing the detected process fault.
|AIChE Member Credits||0.5|
|AIChE Fuels and Petrochemicals Division Members||Free|
|AIChE Graduate Student Members||Free|
|AIChE Undergraduate Student Members||Free|