Research On Improved Hazardous Materials Transport Risk Assessment Model Based On Accident Simulation

With the development of China’s chemical industry,the use of hazardous materials in national life is increasing.Hazardous materials are regarded as a kind of special goods because of their inflammable,explosive,radiation,toxic and other properties,and leakage often results in different types of serious accident consequences.As the leading step of hazardous materials distribution,risk assessment should be comprehensive,scientific,reasonable and compliant to ensure the authenticity and effectiveness of data.However,in the process of risk assessment,the biggest problem is the influence of subjectivity on risk assessment.A single risk assessment model is often one-sided and cannot fully describe the objective risks of alternative sections.Only by risk assessment on sections from multiple perspectives can the transportation risks of sections be truly described.Therefore,it is necessary to conduct in-depth research on the risk assessment model from multiple perspectives,which has positive significance for reducing transportation risks and improving transportation safety.Firstly,this paper collected 2,047 cases of hazardous materials road transport accidents,and analyzed the rules from the time distribution,seasonal distribution and cause distribution of accidents to determine the influencing factors of accidents.The event tree and fault tree in the risk analysis technology are used to study the accident evolution law in detail.First,the event tree is used to analyze and construct the accident scene of explosion,fire and toxic gas diffusion of two types of hazardous materials,and the accident consequences are determined as BLEVE,jet fire and gas diffusion.On this basis,fault tree analysis is used to construct fault trees of BLEVE,jet fire and gas diffusion respectively.Based on fault tree analysis,evolution probabilities of three accident types after leakage accidents are calculated,providing data support for subsequent risk evaluation models.Secondly,the risk value is defined based on the traditional risk assessment model and expressed as the product of accident probability and accident consequence.In terms of accident probability,the multivariate nonlinear regression method was used to calculate China’s predicted freight vehicle accident rate,and the adjustment coefficient of Sino-US freight vehicle accident rate was constructed.The current Chinese freight vehicle accident rate was calculated by referring to the US freight vehicle accident rate,and the conditional leakage probability and correction coefficient were introduced to build the accident probability model.In terms of accident consequences,the dead population was taken as the index,and the dead area and death rate of three types of accidents were determined by ALOHA simulation method of liquid ammonia leakage.The road population was calculated based on the spetiondensity linear model,and the population density under the road was defined as the average population density of the area where the road section was located.The evolutionary probability of three accident types was introduced.The cumulative death toll of the three types of accidents is used as an indicator of accident consequences.In addition,the social risk model was introduced to verify the risk acceptance range of alternative road segments based on F-N curve.Then,the accident type and accident consequence are taken as the main indexes of the risk classification evaluation model,and the risk classification evaluation model is constructed.In terms of accident type prediction,the accident type was taken as the dependent variable and five independent variables,including flammability,rescue time,weather factor,package type and driving speed,were selected.Ordered Logistic regression analysis was carried out on the collected accident data of 2047 cases,and regression function was constructed to predict accident type.In terms of accident consequences,the death toll was divided according to the accident grades in the Regulations on Production Safety Accident Reporting,Investigation and Treatment.ALOHA was used to simulate the accident range and death rate of liquid ammonia of 1～20t transportation volume,and the results were converted into the accident consequence grade matrix of traffic volume and population density.Then,the K-means clustering algorithm is used to conduct cluster analysis on the traffic volume and population density respectively,and the traffic volume and population density are classified,and then the three-dimensional risk grade matrix for predicting accident types,traffic volume and population density is obtained.After the model is constructed,the transportation route risk calculation function is constructed based on triangular fuzzy number.Finally,a chemical plant in Lanzhou was taken as an example to analyze the two models.The results show that the risk values reflected by the two models are consistent with the population distribution and conditions of each section in the example.Compared with the traditional risk assessment model,the risk classification assessment model is simpler to operate and has certain practicability.