基于空间故障树理论的系统故障定位方法研究

Research on the Method of System Fault Location Based on Space Fault Tree

为了系统发生不同类型故障后快速定位可能引起该故障的系统元件,通过分析系统结构和元件故障概率分布,以及系统在不同工作环境中发生各类型故障的统计数量,提出基于空间故障树（Space Fault Tree,SFT）理论的系统故障定位方法。该方法使用SFT概念得到系统内部结构及元件的故障概率矩阵P（Xi）,分析元件X_i故障对于所在割集S_j及系统T故障的贡献度,结合系统故障次数统计矩阵Γ（mq）,最终得到元件X（1～I）与故障m（1～Q）的相关度矩阵。这个矩阵可反映出对于任意系统故障m_q与故障元件X（1～I）的相关性排序、对应的割集、及保证结论正确的可能性,还可优化系统故障分类。实例研究表明：方法可确定各故障的至因故障元件,并根据可能性进行排序,排序靠前的元件组合正是系统的割集,这从侧面也说明了方法正确性。

In order to determine the components that cause the different types of system fault fast after the fault, through the analysis of system structure and component failure probability distribution, as well as the statistical quantity of various types of fault happened in different working conditions, the method of system fault location is proposed based on Space Fault Tree. The method uses the SFT concept to obtain the system structure and the failure probability matrix P（Xi） of the components. The contribution of failure of components Xi to the fault of system T and cut set Sj that contains the component are analysis. Combining with the system failure frequency statistics matrix F（mq）, finally the correlation matrix of the components X（1～I） and fault m（1～Q） is obtained. This matrix can reflect the correlation order of fault component for any system failure, the corresponding cut set, the possibility of ensure correct conclusion, and optimizing the system fault classification. Case study shows that this method can determine the fault components lead to the fault system and according to the possibilities for sorting, the components in the top sort is cut set of the system, which explains the method validity.