Variability of bridge deck vibrations due to traffic loading

F. Besseling ^{1, 2}, P.T. Kockelkorn ¹, J. de Bruijn ², E. Lourens ¹

¹ Delft University of Technology, the Netherlands
² Witteveen+Bos Consulting Engineers

Bridges are key objects in infrastructure networks. Many large bridges in Western Europe were constructed in the 50s and 60s of the past century, and are therefore beyond their design lifetime. Moreover, increased traffic loads often exceed the design capacity. Structural health monitoring can be employed for damage detection and risk management for structures of which the capacity cannot be proven to meet the Eurocode design requirements. When employing vibration-based damage detection methods for monitoring the structural health of bridges, it is often possible to increase the data features' sensitivity to damage by focusing on local as opposed to global vibrations. This increased sensitivity, however, comes at a cost: by moving towards the higher frequency ranges and more local behaviour, the effects of operational and environmental variability on the data become increasingly pronounced. Some form of data normalization needs to be employed in order to reduce this variability in the damage-sensitive features extracted from the monitoring data. To effectively design such normalization strategies, a better understanding of the nature of operational and environmental variability is necessary. This paper presents the results of a study focussing on operational variability, based on a long-term monitoring campaign at the Haringvliet bridge, a steel box-girder bridge in the Netherlands. An analytical model is constructed to study the variations in the measured response, by comparing model vibrations to measured vibrations. The influence of vehicle configuration and speed on the vibrations are demonstrated.

Key words: SHM, bridge monitoring, vibration-based, operational variability