Reliability assessment and maintenance optimization of a load-sharing system considering dependent failure rates and component heterogeneity

In this paper, we consider the reliability and maintenance assessment of a load-sharing system. The system consists of multiple heterogeneous components. We assume that an increase in load leads to the increase of the failure rate of each working component. Two system configurations are studied: the K -out-of- N systems and systems with critical components. The system reliability measures such as the reliability function, the conditional reliability function and the remaining useful lifetime of systems are derived. By assuming that the system failures are hidden, a periodical inspection is proposed, upon which, the decision-maker has the flexibility to do nothing and wait for the next inspection epoch, or to replace the failed components. The expected discounted maintenance cost in the long-time horizon is taken as the objective function. The problem is casted into the Markov decision process. We utilize the value iteration algorithm to solve the optimization problem. A numerical example is presented to show the applicability of the proposed model. The impact of the load-sharing has been comprehensively investigated through sensitivity analysis. It is seen that load-sharing shows significant impact on system reliability and maintenance costs. It can provide theoretical reference to the decision-maker in developing maintenance policies and controlling maintenance budget.