Identification of pedestrian loads on a very flexible footbridge

Abstract

The pedestrian loading on civil engineering structures is usually determined by using equivalent load models or by directly measuring the imposed forces on a rigid floor. Direct measurements are however costly and often only feasible in the laboratory. Additionally, a pedestrian that moves on a flexible bridge tends to adapt to the bridge motion and the load models currently available are therefore no longer valid. The objective of this thesis is to use an inverse dynamic technique to identify the actual loading based on measured vibrations.

The acceleration measurements are carried out on the Geierlay footbridge in Germany. These are then analysed to determine the modal properties of the structure, such as the natural frequencies and the damping ratios.

A joint input-state estimation algorithm is used to identify the pedestrian load, knowing the output accelerations. Before applying this inverse force identification technique to the real-world example, it is validated by two academic test cases. The first one is a 7m long, simply-supported concrete slab, such as in K. Van Nimmen et al. 'Inverse identification of pedestrian-induced loads' (ISMA, 2016), where the modal properties are already given as well as the resulting modal load. The influence of different parameters is also analysed. The second example is the extension of the former, by using the Geierlays length and its modal properties.

The method is then applied to the Geierlay footbridge using the measured accelerations. Finally, the results are compared to existing load models with the goal to propose improvements.