Alarm Root-Cause Identification for Petrochemical Process System Based on Fluctuation Correlation Analysis

摘要

Current alarm system for the chemical plant mainly concerns the single variable alarm. However, in the actual operation, operators hope to obtain information of other variables that affect the alarm variable so as to judge the problems of the device timely. In order to solve this demand, this paper proposes a new method for root-cause analysis Fluctuation Correlation Analysis FCA. By effectively using operating data of devices this method accurately depicts correlation between independent variables, and has low computation complexity, which just meets the need of alarm system for computing speed. We applied FCA to the on-line monitoring system, and it can provide alarm information and operation instruction in time.