考虑阶段备份的多态多阶段任务系统可靠性建模方法

近些年来,由于在航空航天、分布式计算以及核电系统等系统中广泛的应用,多阶段任务系统(PMS)可靠性理论与方法得到了广泛的关注与研究。目前大多数PMS可靠性的相关研究都假设系统的多个任务阶段依次出现,且任一阶段失效都导致整个系统失效。但是在现代航天系统,尤其是载人航天系统,为了使整个系统的任务可靠性更高,经常会采用任务备份的方法来提高系统的任务可靠性。因此,本论文提出一种基于多态多值决策图的方法,通过增加随机变量的方法对单元经历的阶段进行统计,降低所建立的系统模型大小,提高建模效率,并以某型航天器推进子系统为例对建模方法进行说明。

Reliability modeling of aerospace multistate phased mission systems with phase backups

In recent years, the reliability of phased mission systems (PMSs) has attracted much attention, due to their wide application in the aerospace industry, distributed computing and nuclear power systems. Most of the existing research on the PMSs assumed that all the missions are executed sequentially, and the system failed if any one of these missions failed. However, for many PMSs, especially in the aerospace industry, the backup phases are designed for higher mission reliability. To deal with this problem, A multistate multi-valued decision diagram-based method is proposed to model the multistate PMSs with backup phases. By adding an extra variable to the phases that each component passes by, the system model scale and the model efficiency could be improved. Then, based on the system modeling result, the system structure is optimized by an adaptive particle optimization method.