Parallel coordinates plotting as a method in process control hazard identification

Abnormal events in process plants waste billions of dollars annually. Taking advantage of plant operators' perceptual capabilities in the hazard identification process could enhance the aiding provided by other decision support systems. Parallel coordinate plotting allows the visualization of multidimensional data by making apparent geometric patterns that exist in the data sets. This work investigates the use of parallel coordinate plotting in anomalous process control applications. Approaches utilizing the software packages CVE and CPM as decision support tools are explored. These approaches with CVE and CPM are carried out in a case study on historical plant processing data. CVE is used to identify normal plant operating conditions and construct a “best operating zone” that can be loaded into CPM. CPM can then read in real-time or historical data and indicate (i.e, generate alarms) if the plant is leaving normal operating conditions and also recommend mitigation strategies to return the plant to normal operating conditions. CPM's alarms are evaluated as an effective method for determining the onset of a hazard event. Advantages and limitations of CVE and CPM as a decision support are identified through the case study. The results are discussed in the context of the static case study and the need for simulations is highlighted.