In recent years,with the rapid increase in the use of hazardous chemicals,the accident rate of hazardous chemicals in road transportation also shows an upward trend and such accidents often leads to serious accident consequences.To study the dynamic risk change law of road transport accidents of hazardous chemicals,A Bayesian network model was modified and verified and the historical data from 2017 to 2021 was used to conduct machine learning.According to the data of driving fatigue probability changes over time and using the Markov state transition probability matrix calculation method the state transition probability matrix of dynamic node"driver behavior"was obtained,for the dynamic risk prediction model of road transportation of hazardous-chemicals based on Dynamic Bayesian Network(DBN)was established and induced.The study shows that the accuracy of the improved model for the prediction results of the accident consequence nodes is higher or equal to eighty percent,indicating that the model is acceptable.Within three hours of driving,the probability of the driver's"fatigue driving"increases over time,but the rate of increase decreases.In the most common situation,with the nonlinear increase of the driver's probability of"fatigue driving"over time,the probability of"vehicle-rollover"and"crash"accidents increases significantly,which leads to a certain increase of the occurrence probability of the consequences of"leakage"accident.The increase in the probability of"fatigue driving"will lead to an increase in the probability of"casualty"accidents,that is,the increase in fatigue degree aggravates the severity of the accident.In addition,the changing trend of the occurrence probability of"vehicle-rollover","crash","leakage"and"casualty"accidents within the first three hours of driving are consistent with the changing trend of the occurrence probability of"fatigued driving".
维普
万方数据