Signed directed graph-based hierarchical modelling and fault propagation analysis for large-scale systems

摘要

The signed directed graph (SDG) model can be considered as a qualitative model to describe the variables and their cause-effect relations in a continuous process. Such models can allow one to obtain the fault propagation paths using the method of graph search. In this way, the authors can use SDGs to model and analyse the propagation of faults in large-scale industrial systems. However, with increasing system scales, the time requirements of a graph search method would be too onerous. This can be alleviated by transforming a single-layer SDG model into a hierarchical model to improve search efficiency. The hierarchical model would be composed of three layers: the top layer would consist of independent sub-systems; the middle layer would have control systems configuration and the bottom layer would have all the variables. The possible root causes of faults can then be searched in this model, layer by layer according to the initial response of the system. The efficacy of the proposed approach is illustrated by application to a four-tank system and a generator system in a power plant. The methodology presented here can also be used in process hazard analysis.