Study On Accident Analysis And Early Warning Technology Based On Abstract Fault Tree

Historical experience shows that the establishment of an effective accident early warning system can improve the emergency response capacity of the accident.It plays an important role to improve the safety of chemical production,avoid the occurrence of major accidents and reduce the losses caused by accidents.When a local failure occurs in the system,it can detect and identify the accident correctly in the first time,Through the analysis of the accident state tracking and the active intervention control,it is necessary to avoid the following accidents or reduce the loss of the accident.Based on the historical accident case,this paper combines the fault tree analysis method and the accident scenario analysis method to establish the accident analysis and early warning system for the accident scene matching in the chemical industry field.Scenario analysis is an effective mechanism to analyze the state of the accident and the evolution trend.However,due to lack of suitable model the early warning mechanism based on the situation analysis method is difficult to be used in the actual accident early warning method.Fault tree analysis is a kind of common methods of safety analysis,because of its intuitive,image,easy to analysis,is widely applied in all kinds of safety engineering accident mechanism analysis and is widely used in the late accident investigation and analysis.But this model for a specific case is difficult to support the early warning mechanism analysis of evolutionary accident.On the basis of studying a large number of accident cases,this paper puts forward the concept of abstract fault tree of abstract and comprehensive accident,and designs the synthetic method of abstract fault tree for case fault tree,to maximize the preservation of events and the evolutionary mechanism between events,It can effectively support "context-response" type accident analysis and early warning method.Around the content of appeal,the following research is carried out.(1)According to the characteristics of chemical accident case scenario,we put forward the concept of abstract fault tree based on ontology theory and designed an abstract fault tree synthesis method for case fault tree.and defined the synthesis rules and algorithms of abstract fault tree,As the principle of "context-response" accident analysis and early warning method;(2)Based on abstract fault tree,we propose an accident scenario analysis and early warning mechanism for abstract fault tree and its extended cutting set,which can quickly complete the identification and warning of the accident.(3)Based on the concept of importance degree of fault tree diagnosis,we propose an optimization method to calculate the importance of diagnosis,which can balance the relationship between probability and importance,In the field of accident early warning.(4)Based on the principle of search engine,we designed a quick accident scene collection and sorting method.In addition,we designed the algorithm of scene matching based on machine learning method.Through the research content of the appeal,we preliminarily designed and implemented the accident analysis and early warning system based on abstract fault tree.Through some examples show that this kind of accident analysis and early warning technology,It can not only quickly complete the identification and early warning of the accident,but also make a good use of valuable historical case data to reduce the probability of accidents and the loss caused by accident.