Evaluation of TMD Performance in Footbridges Using Human Walking Probabilistic Models

Abstract

Footbridges are generally slender and lightweight structures with low stiffness, designed to support dynamic loads generated by crowds. Therefore, these structures are exposed to vibration problems related to the resonance of human walking step frequencies and the lower vibration modes. To mitigate these problems, one of the most applied corrective strategies is the installation of tuned mass damper (TMD) systems that aim at the vibration reduction of the footbridge’s dominant mode. A fundamental matter in both the footbridge and the TMD design is the pedestrian load modelling, generally considered as a deterministically moving force or a biodynamic model. However, as human gait is a random process, the deterministic models can lead to non-realistic results, directly affecting the TMD system efficiency. In contrast, the use of probabilistic distributions to simulate the human walk randomness can lead to more reliable time series predictions. In this paper, a random walk (RW) algorithm is developed and applied to simulate different crowd scenarios using a simplified plane model of a coupled human-structure-TMD system. In each scenario, the TMD efficiency in reducing the vibration amplitudes is assessed. Results highlight the importance of considering the walking randomness and pedestrians’ dynamic properties in the TMD design.