Affiliations: School of Engineering, University of Warwick, Warwick, UK | Department of Civil and Structural Engineering,
University of Sheffield, Sheffield, UK
Note: [] Corresponding author: Dr. Stana Živanović,
University of Warwick, School of Engineering, Coventry CV4 7AL, UK. Tel.: +44
24 765 28392; Fax: +44 24 764 18922; E-mail: s.zivanovic@warwick.ac.uk
Abstract: Due to their slenderness, many modern footbridges may vibrate
significantly under pedestrian traffic. Consequently, the vibration
serviceability of these structures under human-induced dynamic loading is
becoming their governing design criterion. Many current vibration
serviceability design guidelines, concerned with prediction of the vibration in
the vertical direction, estimate a single response level that corresponds to an
"average" person crossing the bridge with the step frequency that matches a
footbridge natural frequency. However, different pedestrians have different
dynamic excitation potential, and therefore could generate significantly
different vibration response of the bridge structure. This paper aims to
quantify this potential by estimating the range of structural vibrations (in
the vertical direction) that could be induced by different individuals and the
probability of occurrence of any particular vibration level. This is done by
introducing the inter- and intra-subject variability in the walking force
modelling. The former term refers to inability of a pedestrian to induce an
exactly the same force with each step while the latter refers to different
forces (in terms of their magnitude, frequency and crossing speed) induced by
different people. Both types of variability are modelled using the appropriate
probability density functions. The probability distributions were then
implemented into a framework procedure for vibration response prediction under
a single person excitation. Instead of a single response value obtained using
currently available design guidelines, this new framework yields a range of
possible acceleration responses induced by different people and a distribution
function for these responses. The acceleration ranges estimated are then
compared with experimental data from two real-life footbridges. The substantial
differences in the dynamic response induced by different people are obtained in
both the numerical and the experimental results presented. These results
therefore confirm huge variability in different people's dynamic potential to
excite the structure. The proposed approach for quantifying this variability
could be used as a sound basis for development of new probability-based
vibration serviceability assessment procedures for pedestrian bridges.
