Intelligent alarm management

Abstract

An ergonomic problem for the plant operators has appeared in the modern electronic control systems, in which configure an alarm is very easy. We present a methodology and an intelligent software tool to manage alarms and make early fault detection and diagnosis in industrial processes, integrating three techniques to detect and diagnose faults. The three techniques use available information in industrial environments: The alarms of the electronic control system; the fault knowledgebase of the process, formulated in terms of rules; and a simplified model used to detect disturbances in the process. A prototype in a Fluid Catalytic Cracking process is shown.