A MODEL FOR ASSESSING COLLISION RISK ON AUTOMATIC LEVEL CROSSINGS

Mohamed Ghazel

Last modified: 2017-02-28

Abstract


Railway engineers always point out Level Crossings (LC) as a one of the most critical points in railway networks from the safety point of view. Statistics show that more than 300 people are killed every year in Europe in more than 1200 accidents occurring at LCs. In this article, we carry out a comparative study involving two main types of Automatic Protection Systems (APS), the first using a pair of half-barriers and the second with four half-barriers. Each of these two automatic protection systems has some relative advantages and some drawbacks. In the literature and in railway and road guidance documents, some recommendations are given in order to make the choice between these systems. However, these recommendations are mainly based on a qualitative analysis.
Here, we suggest some behavioural models that can be used as a basis to assess the collision risk quantitatively on both LC configurations. These Stochastic Petri net (SPN) models describe the global dynamics within the LC area while taking into account both technical aspects and human behaviour. Our models are parameterizable and allow a realistic representation of the dynamics through the LC. Moreover, their simulation gives us a clear idea about the risk level according to various features of the dynamics within the level crossing area.

Keywords


Railway safety, level crossings, train-car collision, risk-assessment, stochastic Petri net, Simulation

Full Text: PDF