A Monte Carlo Method for Estimating Reliability Parameters of a Complex Repairable Technical System With Inter-Component Dependencies

A general model of a complex technical system is considered. The system is built of multiple two-state components which can be either operable or failed. The functioning of a component is influenced in a specific way by the other components´ states, thus the components are not mutually s-independent. Failures occur randomly, and are handled by several repair teams. If a repair team is available when a failure occurs, then the repair (or replacement) is started immediately; otherwise the component waits in the repair queue. It is assumed that each component´s time to failure is exponentially distributed, and the failure intensity depends on the other components´ states. No assumption is made about the components´ repair time distributions. The model´s complexity makes it impossible to analytically compute the parameters of the system´s failure-repair process. For this reason, the sought parameters are evaluated using a combination of Monte Carlo simulation and statistical estimation. Finding the confidence interval is a non-trivial task. The first part of the paper is theoretical; the conditions under which the failure-repair process is recurrent are given, and the confidence intervals for the sought parameters are defined. The second part has an applicable character; a commodity transport network is considered as an exemplary system with inter-component dependencies, and the algorithm estimating its reliability parameters is presented with some illustrative examples.