Reliability Modeling for Complex System Based on Failure Network Digraph

The traditional system reliability model for complex system exhibits deviation due to the overlook of failure correlation directionality among components. In order to rectify this problem, a new reliability modeling method for complex system is proposed based on failure network digraph technique. First, the failure network digraph is built with related failure mechanism analysis, and the component reliability models are built based on mean rank order. Then, taking the component reliability models as the marginal distribution, the system joint reliability function is developed with the introduction of the Copula function. Third, the parameter of Copula function is estimated with the aid of Kendall rank correlation coefficient obtained by samples. Finally, the method is applied to a machining center, the result shows that the new Copula function model considering failure correlation directionality is more rational and effective compared with the traditional Copula function model, which provides a theoretical basis for reliability evaluation and allocation of complex system.