SFEP文本因果关系提取及其与SFN转化研究

Causality extraction of SFEP text and its conversion to SFN

为将系统故障演化过程(system fault evolution process,SFEP)的文本描述转化为空间故障网络(space fault network,SFN)结构,用于故障分析,本文提出SFEP文本因果关系提取方法,及其与SFN基本结构的转化方法。首先给出SFEP中事件的几种典型因果关系。随后提出因果关系与SFN基本结构的转化流程。本文方法围绕着关键字和因果关系组模式展开,通过模型的不断学习补充和丰富关键字和组模式。最终使方法具备将SFEP文本转化为SFN结构的能力。以飞机起落架故障发生过程文本为例进行了应用,实验结果表明该方法可用于SFEP文本中的因果关系分析,并得到了理想的SFN。完善的关键字和组模式有利于使用计算机智能处理SFEP的SFN。

To transform the text description of the system fault evolution process(SFEP)into a space fault network(SFN)for fault analyses,we propose a method for extracting the text causality of SFEP and its transformation with the basic structure of SFN.First,several typical causal relationships of events in SFEP are given.Then,the transformation process of causality and SFN basic structure is proposed.The method focuses on the key words and causal group models;the keywords and group models are enriched and supplemented through continuous model learning.Finally,the method transforms SFEP text into an SFN structure.The method is applied to the text processing of aircraft landing gear failure.The findings show that the method can be used to analyze causality in SFEP text and obtain an ideal SFN.Perfect keywords and group models are beneficial to the intelligent processing of SFEP text to SFN using computers.