Railway bridges have a long service life - in Germany, for example, the average is 122 years. During this time, vehicles are constantly evolving, leading to innovative axle configurations and higher speeds. This leads to loads for which the bridges were not originally designed, which can lead to resonant structural responses. A key parameter for calculating and evaluating the dynamic behaviour of bridge structures, is the resonance frequency. Due to the large number of bridges in the network, conventional methods using sensors on the structure would be very laborious. A cost-effective alternative would be drive-by monitoring, using sensors on a passing train. In railway bridge contexts, drive-by monitoring to identify frequencies faces challenges from short spans and high speeds, leading to brief contact periods. The brief periods cause a low resolution of the signals in the frequency domain. An approach to overcome this issue is the use of resonance curve-based drive-by monitoring. The term resonance curve describes the maximum structural responses as a function of speed. In conjunction with the known axle configuration of the train, the resonance frequency can be determined from the resonance curve. In this paper, data from two field tests are analysed, in which synchronised acceleration measurements were conducted on the main girders of the bridges and the axle boxes of the trains. An ICE 4 was used on a bridge with a 19.5-metre span, and an ICE TD on a bridge spanning 16.4 metres. The analyses show that frequency identification using indirect resonance curves is feasible under real operating conditions. Thus, this approach enables the determination of a key parameter essential for calibrating structural models. Furthermore, this method enables the avoidance of resonance crossings by speed adaptation, either by acceleration or deceleration. This strategy has the potential to significantly increase the service life of bridges.
Read full abstract