为保证空间相机软件系统可靠安全地运行,多采用定量化故障树方法对其进行可靠性分析。由于获取故障知识及确定故障树结构有一定难度,提出了一种基于粗糙集的软件故障树建立方法。针对软件故障原因错综复杂,无法获取大量有效数据的情况,给出一种采用Vague集方法对底事件进行建模,从而计算底事件相对重要度和顶事件发生概率的定量分析方法,根据分析结果指导软件的可靠性设计并确定测试重点。实验结果表明:使用提出的故障树分析方法,有效地指导了设计人员通过容错差错和避错等设计措施提高系统的可靠性,解决了由于软件部件可靠性指标很难精确量化,无法进行定量分析的问题。
Quantitative Fault Tree Analysis (FTA) is very effective in analyzing the reliability and safety of space camera software system. Aiming at the difficuly to get information of system faults and construct of fault tree, a method establishing software fault tree based on Rough set is put for- ward. Because software failure cause is anfractuous and availahle data are hard to get, a modelling method for bottom event is proposed based on Vague set to calculate the relative significance of bottom event and the occurrence probability of top event. Then, analyzed result from the model is used to guide design of software reliability. Experimental results show that the new method of software FTA can help designer improve system reliability by fault tolerance and error design. The results also show that the new method makes software quantitative analysis possible and optimizes the precision of soft- ware measurement.