Evaluation of Component Reliability Importance for a Multi-component System Using Cross-evaluation based DEA

Nowadays, modern systems have been becoming more complicated because of higher reliability requirements. Redundancy allocation, hot swappable, on-line repairable, multi-stage interconnection designs have been seen increasingly in practice for improving system reliability in system design. In order to meet reliability requirement through these methods efficiently, we should first identify the prioritization of all the components in the system. Some components might be over maintained while others might need more attention. Reliability importance indices are often used to measure the importance of the components to the system. Reliability importance is to identify the weakness in a system and to quantify the impact of the critical component failures. Generally, reliability importance is a function of operation time, of failure and repair characteristics for all the components in the system. Hence, component reliability importance indices are useful for prioritization of maintenance activities related to improvement of system reliability. For that reason, component reliability importance is often considered to determine which components are more important to system reliability improvement or more critical to system failure [1]. In the reliability literature, a number of generally known reliability importance indices already exist such as Birnbaum component importance, structural importance, reliability critical importance, failure criticality importance, restoration critical importance and operational critical importance [2][3]. Zhu et al. [4] analyzed Birnbaum importance measure for consecutive k-out-of-n systems. Also, structural importance measure has been used to measure for the topographic importance of the logic position of components in the system [5]. Hong and Lie [6] introduced a joint reliability importance of two edges in an undirected network. The phase mission in which the system logic may be different between each phase is measure [7]. In this paper, we propose a new evaluation approach of component reliability importance. The five importance factors (failure rate, repair time, repair cost, criticality and safety) are considered for ranking the components in the system. Cross-evaluation-based data environment analysis (CE-DEA) model is used to calculate reliability importance indices of all the components in the system.