Due to the low level of engineering management in China,high complexity of the environment,dynamic construction technology and other characteristics,resulting in frequent safety accidents,especially high fall.After an accident occurs,the investigation team will record key information such as the accident’s occurrence,cause,nature and response measures in the form of accident case reports.Accident case reports contain rich safety knowledge and expert experience,but the reuse rate of these cases is low and excavation is not sufficient.In this paper,239 cases of falling accidents from high places from 2013 to 2019 were studied,with a view to finding out the causes and causative mechanisms of accidents,and providing guidance for scientific prevention and control of safety accidents.This paper mainly consists of three core parts: accident cause analysis,safety knowledge base construction and risk identification,case-based reasoning strategy and method design.In the part of accident causes,the key causes and occurrence mechanism of the accident are analyzed by means of text mining and case analysis,and the hierarchical structure of accident causes is constructed by combining 4M theory.In the construction part of the accident knowledge base,the ontology structure of safety knowledge is designed based on the causative theory,and the numerical attributes,object attributes and relation attributes of the knowledge source and ontology are introduced in detail.The parameterization rules are used to identify potential risks,which plays a role in accident prevention.In case-based reasoning(CBR),the similarity of the accident scene,the accident cause similarity and the accident loss similarity are calculated,and an attribute weighting method considering the accident cause level is proposed,and the measures of similar cases are recommended by the idea of nearest neighbor for the reference of the accidents that have happened,which plays a role in the accident control.Research results show that the fall is the key cause of the illegal violation of the workers work,insufficient safety awareness,and regulatory check does not reach the designated position,in the process of operation of the workers’ unsafe behavior and unsafe state of object,such as the mistake is the surface of the cause of the accident,the complexity,the complexity of technology,environment and management complexity is the deep reason of the accident.The results of case analysis show that the potential risks of parameterized entities can be identified by rules,and the corresponding measures of similar cases inferred have practical reference value for the accidents that have occurred,but there are also some limitations.This paper has a certain theoretical and practical significance in the aspect of intelligent response to safety accidents. |