| Shield method has been widely used in tunnel construction,which is the preferred technology for the construction of urban underground infrastructure,subway and pipeline gallery,but it is accompanied by an increasing number of failures.Current risk identification methods have missing risk factors,which indicates that people do not have a comprehensive understanding of failure.Failure will occur again if the cause of failure can not be identified accurately.Therefore,failure has not been effectively utilized,resulting in "information overload" but "lack of knowledge".The rise of BIM technology provides a tool for knowledge acquisition and expression combined with engineering model.This paper focuses on the reverse process from the result to the cause,manages failure knowledge through BIM,realizes the reverse identification and reasoning of risk factors,and visualizes in BIM model.The knowledge management function of BIM can realize the input of failure data and the output of failure knowledge.Through the feature information of current failure in BIM,managers can retrieve the constructed failure knowledge base based on ontology semantic similarity,and extract similar cases.Thus,identify common risk factors by text analysis,and add them to the existing risk factors system to make qualitative and quantitative analysis of the causes for the current failure.Firstly,fault trees of current failure types are generated by failure knowledge base;secondly,the probability of basic events is determined by the fuzzy reasoning mechanism based on cloud model,and transformed into Bayesian network;finally,Multi-Evidence reverse reasoning is carried out through the Bayesian network to find out the causes of failure.The case study of cutterhead wear in Xiamen Metro Line 3 shows that the risk factor of "quartz content" can be excavated from several similar failure cases of cutterhead wear,risk identification based on failure knowledge is a supplement to the existing methods,and the weight of common cause analysis should be increased.Based on this,the results of reverse identification of the current failure reasons show that there are deviations in the expert’s judgment of the cause probability of cutterhead wear,and the objective method is closer to the facts.At the same time,Multi-Evidence reverse reasoning based on Bayesian network provides guidance and direction for rapid identification of failure causes.BIM has the advantage of three-dimensional visualization,which stores failures on BIM and presents them to project participants,so that failure knowledge can be easily understood.This paper can expand the role of BIM in knowledge management,achieve effective management and application of failure,and constantly enrich people’s understanding of failure. |