系统故障因果关系分析的智能驱动方式研究

Intelligent analysis of system fault data and fault causal relationship

为适应未来智能环境和安全领域的故障数据分析需求,本文提出系统故障因果关系分析思想。论述了通过数理统计方法分析系统故障数据存在的问题,研究了系统故障的相关性和关联性,前者基于故障数据反应故障表象;后者基于故障概念反应故障本质。将智能情况下的故障因果分析划分为4个层次,包括数据驱动、因素驱动、数据-因素驱动、数据-因素-假设驱动。该方法的特点是获得广泛的故障因果关系,深入了解因果关系,两者兼顾和更接近于人的思维。4种驱动对故障因果分析的能力依次上升,可为安全科学与智能科学结合提供渠道。

To meet the needs of fault data analysis in the future intelligent environment and safety field,we propose the idea of causal relationship analysis of system fault.First,the problems existing in the analysis of system fault data by mathematical statistics are discussed,the correlation and relevance of system fault are studied,showing that the former is based on faulty data,reflecting the fault representation;the latter is based on fault concept,reflecting fault essence.The fault causal analysis in the intelligent situation is divided into four levels:data driven,factor driven,data-factor driven,and data-factor-hypothesis driven.Their characteristics are to separately obtain and understand a wide range of fault causal relationships,both considering and closer to human thought.The ability of the four drivers to analyze the fault causal relationship increases in turn,it can provide a channel for combining safety science and intelligent science.