Statistical Analysis of Bridge Accidents under Non-Natural Disasters Using Bayesian Networks
This paper introduces a method for analyzing bridge accidents caused by non-natural disasters using Bayesian networks.Accident samples from domestic bridges experiencing non-natural disaster-related accidents during their operational phase were collected and subjected to statistical analysis and preprocessing of their indicators.Preprocessed statistical indicators were used as nodes to establish a Bayesian network model for bridge accidents under non-natural disasters.The model was validated,and through this model,posterior probabilities and sensitivities of various root nodes were inferred,analyzing the impact of various root node factors on the overall bridge damage.Correlations between bridge accident causes were calculated.The research results indicate that the age of the bridge's service has the most significant impact on the overall bridge damage.Factors such as the length of bridge,the construction year of the bridge,whether it is a beam bridge,and factors related to engineering quality,overloading,and collisions with vehicles and ships also influence overall bridge damage.Additionally,the effect of factors such as whether the bridge is made of concrete,is a single-span bridge,crosses water,is a highway bridge,and other causes related to accidents is relatively small and can be disregarded.There is a negative correlation between various bridge accident causes,with particularly high negative correlations between overloading and collisions with vehicles and ships,overloading and engineering quality,and collisions with vehicles and ships and engineering quality.The negative correlations between engineering quality and other causes,overloading and other causes,and collisions with vehicles and ships and other causes are relatively low.
国家科技期刊平台
NSTL
万方数据
