Cause and Effect Analysis in Chemical Processes Utilizing Plant Connectivity Information

Faults and disturbances in chemical process plants provide great challenges to maintenance staff in terms of the time and expenditure spent on trying to locate the root causes. Due to the complex connection of units within any plant and the numerous loops that may be present, the process of managing these faults is a large task and it is a challenge to pin-point exactly the location of the root disturbance. Recent methods and tools have focused on the use of data from the plant in question. However, it is important to note that the physical connectivity of the components within the plant also plays a significant role in the propagation of disturbances. A separate strand of work in the literature has used causal maps and digraphs of the process successfully in fault isolation. Any method alone may succeed in creating a number of spurious solutions. By combining two or more different methods, it is possible to reduce the number of spurious solutions. This article utilizes connectivity information and a data-driven analysis that is able to indicate causal relationships between the measurements in the process in order to generate and test hypotheses about the root cause.