Researcher
Research institution
Champion
Focus team
Topic
Project status
Year ended
2011
Project ID
200902
Why should I care about this project?

This was our first attempt at using DCS data to predict abnormal events.  While unsuccessful, it helped inform subsequent Event Prediction projects.

Abstract

Dr. Ellen Bass of the University of Virginia attempted to mine DCS data to identify the signature of the abnormal events earlier than detection by alarms. Principal component analysis (PCA) was applied to data supplied by member companies. While events were identified earlier with the PCA, the false alarm rate was about 50%. It was decided that the use of a different approach was needed.

Objective
  • Provide advanced warning of pending events so operator can recognize situation and take action in realtime (Detect 10 minutes early)
  • Identify techniques for indicating: (a)What is normal, what is not normal and (b) How process and equipment are operating (trend)
  • Can generalized techniques identify precursors to abnormal situations and system behavior?
Driving questions
  • Can the output of CPM (alarms) be used as an indicator of a hazard event so that it can be used as input to potential decision support tools?
  • Would increase in alarms and knowledge of associated tags provide sufficient warning to the operator?
  • Would CPM advice allow successful mitigation?
  • Can alarms be used as indicator of hazards?
Background

Despite advances in safety interlock systems and instrumentation, situations still arise that operators are unable to accurately assess and result in unplanned equipment outages, lost productivity, and unsafe conditions. Better techniques are needed to illustrate the state of the equipment and process and to provide advanced warning of pending events to the operator. The effort will determine how to identify key indicators, how to present these indicators to operators, and how to incorporate decision aids to assist the operator.

Deliverables
  • Identify techniques for showing what is normal, what is not normal (suggests showing limits or constraints), and where the process and equipment are both operating and heading.
  • Provide one or more mechanisms that can be used for identifying precursors that leads up to an unplanned event.
  • Suggest guidelines for developing displays and navigational aids that can be used to focus operators on where the problem is.
  • To show how decision aids can be incorporated to support operators avoiding or getting the situation.