A Combinatorial Approach to Quantify Stochastic Failure of Complex Component-Based Systems_The Case of an Advanced Railway Level Crossing Surveillance System

There are different approaches to quantify stochastic failures of complex component-based systems. Methods successfully applied such as fault tree analysis, event tree analysis, Markov analysis and failure mode and effect analysis are recommended to certain extend according to their applicability to component-based systems. A combinatorial model of fault tree analysis and Markov analysis is developed in this paper to estimate the safety state of an advanced railway level crossing surveillance system which will be implemented by Taiwan Railways Administration in the future. Based on observations of an existing level crossing system, the combinatorial model is used to determine an instantaneous risk probability function, which is dependent on the system state. The results demonstrate that the advanced railway level crossing surveillance system has a higher safety state probability than the existing one.