空间计算机冗余架构可靠性分析比较

采用冗余架构,是改善和提升空间计算机可靠性的一个重要途径。目前在工程中对不同架构可靠性的对比,尤其是相对复杂的架构,多仅限于形式上的粗略分析,缺乏定量比较。对6种典型空间计算机冗余架构(双机冷备、双机热备、TMR、动态重构TMR、2+1三机、四机)分别建立马尔可夫链数学模型,推导出上述各冗余架构可靠度的计算公式,并以一个虚拟的5年期任务为背景,对这些架构的可靠度进行了仿真计算、分析和比较。仿真结果表明:具有冷备资源的冗余方案可获得相对高的系统可靠性;让故障节点具有从故障中恢复正常功能的能力和短的修复时间,能大幅提升系统可靠性。

Analysis and Comparison of Reliability of Computer Redundancy Architecture in Space

Applying redundant architecture is an important way to improve and promote the reliability of computer in space. The redundant architecture of space computer can be organized in different forms depending on the combination of software and hardware resources. Recently most comparisons among different redundant architectures in engineering are performed only by rough analysis,not in quantitative way,especially for the relatively complex architectures. In this study,six typical redundant architectures(dual-unit with cold standby,dual-unit with hot standby,TMR,re-configurable TMR,2+1 triple-unit,and quad-unit)are investigated,the Markov chain models are established and the formulas of reliability are deduced respectively. Reliability of the architectures is calculated,analyzed and compared based on a virtual five-year space task. The simulation results indicate that a solution with cold standby resource can obtain higher reliability. Moreover,the ability of a fault unit to recover easily and quickly can greatly improve the system reliability.