Risk analysis on level crossings using a causal Bayesian network based approach

Safety is a core issue in railway operation. In particular, Level Crossing (LX) safety is one of the most critical issues that railway stakeholders need to deal with. Accidents at European LXs account for about one-third of the entire railway accidents. They result in more than 300 deaths every year in Europe. However, due to non-deterministic causes, complex operation background and the lack of thorough statistical analysis based on detailed accident/incident data, the risk assessment of LXs remains a challenging task. In this paper, a general approach of Causal Statistic Risk Assessment based on hierarchical Causal Bayesian Networks (CSRACBN) is developed to analyze the various impacting factors which may cause accidents, and identify the factors which contribute most to the accidents at LXs, thus allowing for risk quantification. The detailed statistical analysis is carried out firstly according to the accident/incident database, then the CBN risk model is established based on the statistical results. In order to validate the effectiveness of this approach, we apply the CSRA-CBN approach on the basis of the accident data from SNCF, the French national railway operator. The CBN model allows for quantifying the causal relationships between the risk of accident occurring and the impacting factors considered. Moreover, the hierarchical structured modeling offers interesting benefits in terms of clarity, which makes it possible to highlight the complex factors influenced by a mass of parameters and identify the factors that contribute most to LX accidents. In addition, the main output of such a modeling system is conducive to improving safety at LXs, meanwhile, allowing for efficiently focusing on the effort/budget to make LXs safer.