考虑故障覆盖和切换失效的、带有温备份的冗余系统可靠性研究

Standby sparing technique is usually used to enhance the reliability of complex systems. Typically, there are three types of standby sparing technique, i.e., hot standby sparing, cold standby sparing and warm standby sparing. A warm standby component is partially powered up during standby and is switched to fully operational state only after replacing a failed working component. Fault coverage refers to the successful detection,location,and accommodation of a fault by the system in case it occurs. Even when a system is configured with enough redundancy, a single uncovered component failure may cause the whole system to fail. Only a few existing papers on the reliability of warm standby systems have considered imperfect fault coverage, and these works are restricted to very special cases. Besides, these papers assume that the switch is always perfect to change the warm standby component from standby state to fully operational state. In fact, a warm standby system may fail due to the switching failure. Some recent researchers have incorporated the switching failure into the reliability study of warm standby systems, however, without considering fault coverage. Based on an improved binary decision diagram technique, this study can overcome all these shortcomings. The models proposed can be incorporated into commercial reliability calculating software. Meanwhile, the improved binary decision diagram can be a powerful tool in analyzing the reliability of other complex systems.

复杂系统经常使用备份件增加可靠性。备份件分为热备份、冷备份和温备份。温备份作为备份件时处于部分工作状态，在工作部件失效后才快速通过切换装置取代工作部件进入完全工作状态。故障覆盖,是指系统部件在发生故障时故障被系统完整地检测、定位和恢复。即使在系统有足够的冗余时，系统部件的未覆盖故障也可能会导致系统失效。目前，只有少数几篇研究带有温备份的冗余系统可靠性的文献考虑了故障覆盖，且都是针对一些很特殊的情况。另外，这些研究都假设温备份部件的切换装置是完美的。实际上，一个带有温备份部件的系统可能会由于切换失效而发生故障。最近有些学者在研究带有温备份的冗余系统可靠性时考虑了切换失效，但是又未考虑故障覆盖。本研究通过改进的二分决策图方法研究上述问题，可以大大突破现有研究的局限性。本研究提出的模型可以应用到计算系统可靠性的商业软件中。同时，这种改进的二分决策图方法对其他复杂系统可靠性建模也有很大的借鉴意义。

冗余对于提高系统可靠性是非常重要的手段. 备份件可以分为热备份, 温备份和冷备份. 温备份既能消耗较少的能源, 又能较快地切换成为工作元件, 因此受到越多人关注. 但是, 由于温备份的故障发生率会变, 给带有温备份的冗余系统可靠性研究造成了很大的困难. 国际上的研究多限于只含有一到两个温备份元件的系统, 并且多假设元件故障发生时间服从指数分布. 通过对决策图技术进行改进, 我们取得了很大突破, 研究了含有任意多个温备份元件的冗余系统的可靠性, 并且容许系统元件的故障发生时间服从任意分布. 除了研究系统可靠性之外, 我们还研究了系统的结构优化问题. 另外, 我们还将改进的决策图技术运用到了其他的复杂系统研究, 比较多阶段系统的研究. .发表20多篇SCI论文, 并且受到科技日报报道(2016.12.16) , 对于推广温备份技术, 实现节能减排 , 增加经济效益, 促进环保有着巨大作用.