Footbridge Response Predictions and Their Sensitivity to Stochastic Load Assumptions

Abstract

Knowledge about footbridges response to actions of walking is important in assessments of vibration serviceability. In a number of design codes for footbridges, the vibration serviceability limit state is assessed using a walking load model in which the walking parameters (step frequency, pedestrian mass, dynamic load factor, etc.) are modelled deterministically. This is a simplification of matters, as the walking parameters are basically stochastic as they will be different from one pedestrian to the next. The present paper considers a stochastic approach to modelling the action of pedestrians for predicting footbridge response, which is meaningful, and a step forward. Modelling walking parameters stochastically, however, requires decisions to be made in terms of their statistical distribution and the parameters describing the statistical distribution. The paper investigates the sensitivity of results of computations of bridge response to some of the decisions to be made in this respect. This is a useful approach placing focus on which decisions (and which information) are important for sound estimation of bridge response. The studies involve estimating footbridge responses using Monte-Carlo simulations and focus is on estimating vertical structural response to single person loading.