Enabling comprehensive failure analysis of complex physical system using cognitive map-based approach

摘要

The rapid evolution in technology is leading to complex engineering systems confounding the problems of failure analysis processes. Earlier efforts have concentrated on failure analysis of components but have proven insufficient for analysing the complex physical systems. In this paper, a cognitive map-based approach suggested by Augustine et al. (2012) for system interaction failure analysis is further improved to develop populated network structure considering structural variation. The numerical values to causal arcs are assigned considering qualitative linguistic variables to represent causal relationship between two nodes. The structural variability in cognitive maps due to input from experts or sources is addressed by effectively combining the cognitive map diagrams (networks) and related information received from several sources into single network diagram. A combined adjacency matrix is developed considering credibility weight assigned to each experts. The proposed approach is capable of capturing system interaction failures as well as component failures, which is useful at the early stage of design. A relatively complex physical engineering system of fuel injection is presented as an example to demonstrate the effectiveness of the proposed approach.