Citation

Wang J, Roberts C, Chen L, Zhang Y. Safety Sci. 2016; 86: 228-237.

Copyright

(Copyright © 2016, Elsevier Publishing)

DOI

10.1016/j.ssci.2016.03.003

PMID

unavailable

Abstract

Since 1980, there have been over 100 train collision accidents, resulting in many serious injuries and deaths as well as financial losses. In 2013 alone, there were cases of this type of accident in Switzerland, Argentina, America, Portugal and Russia. There are various causes of train collisions, including overrunning signals at danger, setting the wrong train routes, over-speeding, brake failures, etc. Setting the wrong train routes is regarded as the main cause of collisions.

This paper proposes an approach to eliminate train route setting errors by applying parallel monitoring, which can be used to supervise train route setting and operator operation processes based on virtual image technique. Unlike a real interlocking system, this method can link together the traffic planning, operator operation, interlocking execution, train route occupation and releasing, to achieve the purpose of enhancing system safety and efficiency, whilst monitoring operator operations.

The problems with existing train route setting systems are analyzed and a solution is proposed. The principle and methodology of using parallel monitoring to supervise train route setting based on virtual image technique are described. A designed architecture is illustrated and elaborated for a good understanding of the system. A case study of a Stochastic Petri Net (SPN) model using the proposed parallel monitoring is given, based on which the safety of the system is analyzed.

Language: en