Girder integration strategy: a measure to reduce seismic derailment risk of trains on irregular bridges
Abstract
Under seismic excitation, irregular girder bridges with abrupt changes in section and mass often exhibit large dynamic response differences at key locations, leading to localized track deformation and elevated risks to train running safety. To address this issue, a comprehensive three-dimensional coupled dynamic model including a train, ballasted track, and irregular bridge is developed. A multi-timesteps strong coupling co-simulation strategy based on MATLAB–OpenSees is adopted to ensure both computational accuracy and efficiency, while a cyclic train-running strategy is introduced for long-duration earthquake analysis. To control excessive relative displacements at girder ends, girder–girder limiting devices (GGLDs) with different stiffness values are installed. The results show that running safety risks are mainly concentrated at the embankment–bridge transition and at locations with sudden changes in girder width, where rail deformation is most severe. GGLDs can effectively mitigate these adverse effects, reducing the peak lateral wheel–rail displacement by up to 40% and the fastener displacement by over 90%, thereby converting some derailment cases into safe running conditions. However, the improved running safety is accompanied by increased fixed-bearing demand, indicating a clear tradeoff between train protection and structural force transfer. Further analysis of train-response gradients and running fragility demonstrates that a GGLD stiffness coefficient α in the range of 0.5–0.6 provides the most favorable balance between derailment-risk reduction and structural demand control.
Fields of Study & Subjects
Railway Engineering and Dynamics
Engineering and Material Science Research
Geotechnical Engineering and Underground Structures
Publication details
- Journal / source
- Mechanics Based Design of Structures and Machines
- Publisher
- Informa UK Limited
- Volume
- 54
- Issue
- 1
- Publication date
- 2 January 2026
- Language
- EN