Technical gazette, Vol. 32 No. 6, 2025.
Original scientific paper
https://doi.org/10.17559/TV-20250410002573
Research on Risk Evolution and Scenario Deduction of Railway Emergencies Based on Knowledge Elements and Dynamic Bayesian Networks
Chang Liu
; School of Economics and Management, Beijing Jiaotong University, Haidian, 100044, China
Dan Chang
; School of Economics and Management, Beijing Jiaotong University, Haidian, 100044, China
Daqing Gong
; School of Economics and Management, Beijing Jiaotong University, Haidian, 100044, China
*
* Corresponding author.
Abstract
Railway transportation safety is crucial to the national economy and people's livelihood. Conducting risk evolution analysis and scenario deduction for emergencies is a key to preventing and controlling accidents. This paper proposes a method integrating knowledge elements and dynamic Bayesian networks to realize risk evolution analysis and scenario deduction of railway emergencies: firstly, constructing a knowledge element model of railway emergencies to formally express event elements and their associations; secondly, building a dynamic Bayesian network based on this model to depict the dynamic risk transmission process; thirdly, training the network through parameter learning algorithms combined with historical data and expert knowledge; finally, carrying out scenario deduction based on the trained network, analyzing risk evolution trends under multiple scenarios and proposing prevention and control measures. The case verification of the D2809 train derailment accident shows that this method can effectively simulate the risk evolution process of emergencies and provide scientific decision support for railway safety risk management.
Keywords
dynamic Bayesian networks; knowledge elements; railway safety; risk evolution; scenario deduction
Hrčak ID:
337737
URI
Publication date:
31.10.2025.
Visits: 121 *