A Monte Carlo Simulation-Based Algorithm for a Repairable System in GO Methodology

Abstract

GO methodology is a significant system probability analysis technique. Because the GO model is similar to system schematic diagram, it is easy to establish and often used to evaluate the reliability and safety of complex systems. The GO model can not only describe multi-state and time-sequential characteristics, but also express maintenance behaviors of systems. However, the solutions for repairable GO model through analytical approaches such as the numerical method and methods combined with Markov models or dynamic Bayesian networks have limitations in terms of shared signals, complex dynamic maintenance behaviors, the number of operators, non-exponential failure or repair probability distributions and the size of networks. Based on this, a Monte Carlo simulation-based algorithm for repairable GO model is proposed in this paper. The process of generating state time diagrams and the simulation principle of several typical operators are given and the result of an example validates the algorithm. By this new algorithm, the curves of reliability with time and the instantaneous reliability indices on determined time can be figured out without considering the limitations mentioned above.