Research On Accident Prediction Of Road Dangerous Goods Tanker Transportation Based On Bayesian Network

Dangerous goods are often flammable and explosive,and highly polluting to the environment.In case of accidents in the transportation of dangerous goods,it is likely to cause long-term and huge harm to people's production and living and ecological environment.Study the influencing factors of road transportation accidents of dangerous goods,predict the probability of road transportation accidents of dangerous goods,identify the risks in the transportation,and build a road for dangerous goods It is of great significance for transportation enterprises to formulate effective risk reduction measures.In this paper,the Bayesian network prediction model of road transportation accidents of dangerous goods is constructed by using association rules,interpretive structure model and Bayesian network knowledge,and is verified by accident examples.The main contents of this paper are as follows:Based on the data of more than 1000 road transport accidents of dangerous goods and more than 200 simulated driving tests of road transport tankers in 2013-2018,Apriori algorithm is used to mine the factors affecting accidents,extract and analyze the association rules with the occurrence of accidents as the subsequent item,and determine the unsafe driving Behavior,operation error,vehicle failure(including tank),road type,road alignment,weather,lighting conditions,time period,month and dangerous goods type are the key influencing factors of road transportation accidents of dangerous goods.Using association rules to improve the traditional interpretation structure model,by mining the association rules between the influencing factors of road transport accidents of dangerous goods,combined with expert knowledge to judge the relationship between the factors,so as to establish the structure interpretation model and establish the hierarchical relationship among the influencing factors.Based on the hierarchical relationship among factors and the learning method of Bayesian network structure,the Bayesian network structure of road transport accidents of dangerous goods is established,and the parameters of the network structure are learned by using the special Bayesian network software.Finally,the prediction model of road transport accidents of dangerous goods is constructed,and the accuracy of the model is verified,and the impact of the prediction model on road transport accidents of dangerous goods is verified The sensitivity of factors was sorted and the accident cases were analyzed.