Influence of Cross-Sectional Configuration on the Dynamic Response of High-Speed Railway Bridges

This study investigates the effect of cross-sectional configuration and skewness on the dynamic response of short-span, high-speed railway bridges. Three-dimensional finite element models were developed for the following four 30 m simply supported bridge configurations: I-girder, U-girder, hollow slab and box girder. To enable a controlled comparison, the following parameters were maintained consistently across the investigated cases: span length, deck depth, material properties, support idealisation, track position and HSLM-A1 loading scheme. A single train load was applied along an eccentrically positioned track to activate vertical bending and torsional responses. The first vertical bending frequencies of the straight bridges were closely grouped between 4.63 Hz and 4.67 Hz. However, their speed-dependent responses differed substantially, indicating that the fundamental bending frequency alone is insufficient to explain the dynamic behaviour of bridge decks with different structural configurations. The hollow slab bridge exhibited the lowest response levels over the investigated speed range, whereas the I-girder bridge showed the greatest variation between the centre of the bridge and the position of the loaded track. This behaviour was associated with the proximity of the bending and torsional modes under eccentric loading. As the skew angle increased from 0° to 40°, the hollow slab bridge showed the greatest increase in the first vertical bending frequency (approximately 22.5%), while the I-girder bridge showed only a slight increase (approximately 1.5%). In this study, the resonance formula based on the fundamental bending frequency is used as a reference estimate for straight bridges, rather than as a complete predictor of the response of skew bridge decks. The results show that the effect of skewness depends on the configuration and is governed by the combined effects of modal characteristics, transverse stiffness distribution, torsional response and support geometry. These findings are specific to the adopted moving-load model and provide evidence to inform the selection of an appropriate analytical level in the preliminary dynamic assessment of high-speed railway bridges.