Nowadays, building floors are prone to excessive vibrations induced by human actions, especially when a group of people perform rhythmic activities in a coordinated manner. Resulting effects on the floor occupants vary from perception, discomfort or even panic.
This thesis addresses the development of a spectral modelling approach related to crowd-rhythmic activities, and proposes methods for predicting the response of floors subjected to such load cases. Chapter 2 presents a literature review about the vibration of floors subjected to rhythmic activities. Chapter 3 describes the first experimental campaign carried out on a floor specimen where individuals performed various rhythmic activities while subjected to audible and visual stimulus. A frequency-domain load model characterizing crowd-rhythmic activities is established in Chapter 4, comprising a spectral load model for a single person, combined with coordination factors for multiple individuals. The procedure to identify the crowd model parameters is detailed afterwards and applied to laboratory measurements for four investigated rhythmic activities. The developed models are then validated and their scope extended in Chapter 5, based on experiments realized on a full-scale floor. These crowd load models are then compared with existing models in the literature and used in Chapter 6 to propose design-oriented methods for the evaluation of floor responses due to crowd-rhythmic activities.