Dynamic Impact Factor Evaluation For Short Span Steel Railway Bridges

The Dynamic Impact Factor (DIF) is widely employed to account for the dynamic amplification effect of moving trains on railway bridges. An accurate DIF provides a safe yet economical basis for new railway bridges and improves the safety rating assessment for existing railway bridges. This thesis investigates the accuracy, reliability and the underlying influence factors for DIF relationships currently used for short span steel railway bridges. Full-scale field monitoring exercises are conducted to measure the dynamic responses of two railway bridges during various train passages. The monitoring results indicate that both railway bridges satisfy the live load deflection limits recommended for railway bridges subjected to low-speed trains. Three-dimensional Finite Element (FE) models are developed for the each of railway bridges. The models are verified against the monitoring results. The verification results show that the models accurately predict the actual dynamic response of the railway bridges. A series of sensitivity analysis is performed using the verified FE models. The analyses investigate the effects of variation in New Zealand train and bridge dynamic characteristics on the mid-span DIFs of the monitored railway bridges. The trains are simulated as moving constant forces. The analysis results show that the train speeds have the largest influence on the DIFs of the railway bridges. Numbers and axle distances of carriages have some effects on the DIFs which these effects depend on corresponding locomotive axle distances. Bridge damping ratios have some influences, and the train axle loads have no effect on the DIFs. Over 100 different train arrangements corresponding to combinations possible in New Zealand are simulated and applied to the each of the FE modes. The mid-span DIFs are evaluated numerically as the simulated train passes over the bridges with different speeds. It is found that the DIF formulas in New Zealand railway bridge guidelines overestimate the dynamic effects of moving trains on the railway bridges. This overestimation approaches 4.2 times the bridge evaluated DIFs. Data mining techniques are employed to generate predictive models which estimate the medians of the simulated DIFs. These predictive models provide users with a reliable prediction of the DIFs for designing or assessing the short span steel railway bridges subjected to train passages with speeds up to 150 km/h.