Dr. Thomas Edgar of the University of Texas utilized the data for compressor surge events to identify process faults prior to alarm actuation. Application of parallel coordinate analysis to principal components was expanded by creation convex hulls superimposed on 3D radial plots, allowing compressor surge to be detect two hours prior to the event. Application of the technique to the Eastman Challenge outperformed other published techniques. 3D plots are problematic for continuous monitoring, so a better presentation method is needed.
The proposed work will determine how to make use of data that characterize abnormal conditions to identify and present such indicators to operators, and how to incorporate decision aids to assist the operators in an industrial environment.
- Which data mining visual tools can be implemented to assist operators during process monitoring?
- How to extract meaningful statistical features from normal and abnormal operating data?
- How such features or indicators can be visually represented in a graphical environment?
- Which data features can be used to predict the coming of an undesirable operating condition?
- How to reduce the number of false alarms when predicting the coming of a potential fault?
- What is the probability of successful predicted faults?
- How to incorporate operator decision aids in the proposed methodology?
A significant number of undesirable process events can be prevented by the use of empirical indexes and visualization techniques preconfigured for recorded historian data. The development of such indicators depends on the data features captured during abnormal conditions.
- Identify effective graphical tools to be used for process monitoring and fault detection.
- Suggest guidelines for developing displays and navigational aids that can be used to focus operators on where to find fault causes.
- Provide one or more mechanisms that can be used for identifying precursors that lead up to an unplanned event.
- To show how decision aids can be incorporated to support operators avoiding an abnormal condition.