A Distributed Intelligence System for Improving Fault Diagnostic Performance in Large Scale Chemical Processes

Faults in large-scale chemical plants could occur at a process level although, more often, faults occur at the instrument and equipment level. A failure in an equipment could quickly propagate throughout the process resulting in leaks, fires and explosions causing loss of life, capital invested and production downtime. Attempting to monitor the overall process for identifying such instrument and equipment level failures may be futile as deviations in the process are often lagging indicators by which time the plant safety may be compromised. Hence, we propose a distributed process monitoring system which uses multiple FDI methods/agents capable of monitoring the plant at various sections, levels of granularity (tag level to unit level) and on various operating states. The process is divided into multiple scales and multiple operating states and FDI agents are developed at these scales/states. When multiple FDI agents are used they need to effectively interact with one another. Hence, a Process Ontology is developed to explicitly capture the hierarchy of the process. Since different types of faults at different levels of granularity in the plant could occur, a Fault Ontology is also developed and mapped to the process ontology. The proposed approach, called ENCORE, has been implemented as a multi-agent system and its efficacy is demonstrated on an offshore natural gas production platform.