Supporting plant disturbance analysis by dynamic causal digraphs and propagation look-up tables

This paper presents a method for the analysis of plant disturbances based on the combined monitoring of different but interrelated propagation carriers, specifically material flow-type and information flow-type carriers. The underlying notion of the proposed approach is that certain properties of plant assets, so-called propagation-related factors (PRFs), determine the behavior of disturbances being transported through the process by complex carriers' combinations. Accordingly, the method relies on the systematic evaluation of PRFs and their specific influence on monitored carriers for the discovery of disturbance paths. Such a procedure is executed by a) the event-triggered generation of Dynamic Causal Digraphs (DCDGs), i.e. dynamic causal structures relating plant topology, flow directionality, propagation carriers, and plant asset properties, and b) the systematic consultation of a Propagation Look-Up Table containing PRF-based evaluation functions for the effect estimation of each node in a DCDG. As a result, the method allows for the ranking of discovered disturbance paths and the isolation of faulty plant zones, serving thereby as a basis for automated disturbance analysis and alarm grouping in process facilities.