Identifying Critical Failure-Propagation in Function Models of Complex Systems

Abstract

Complex interconnected systems have high demands on meaningful analysis of the impact of failures on the actual service provision. This includes the study of obvious and high probable events, but also failures that are difficult to anticipate, e.g. due to cascading effects or combined events. This work introduces a framework for failure analysis that enables the exhaustive identification of combined failures with the strongest impact on the functionality of a system. The framework consists of two principal elements: a method for capturing the propagation of failures in complex systems that are represented via function models, and algorithms for solving the identification problem, which is formulated as combinatorial optimization problem. The feasibility of the approach is verified at hand of a function model of an Offshore Wind Farm (OWF). Both algorithms are then applied to the model of an offshore wind farm in order to identify the failure combinations with the strongest impact on the functionality.