Chapter 1 - Bayesian networks for failure analysis of complex systems using different data sources

Abstract

Failure analysis is an important and challenging aspect of the study of complex systems. A system is defined to be consisting of components, subsystems, inputs, and outputs within system boundaries. The inputs provide physical resources and information to the subsystems, which are interacting among each other to produce some outputs. All interactions are assumed to take place within system boundaries. A complex system can be defined as a system structure that is composed of usually a large number of components that have complex interactions. Any failure in performing the required interactions among system components, or any failure in getting the expected output/result, is considered to be contributing to system failure. Thus, analysis of a system with its components is a crucial step in determining the difficulties and complexities that the system will experience at any stage. However, in the real world, performance of both inputs and subsystems is affected by probabilistic uncertainty, and hence, a failure may come with an associated probability. The main goal of this chapter is to evaluate the probability of failure of complex systems, while finding the failure causes using Bayesian Networks (BNs).