Time-domain numerical analysis of single pedestrian random walks on laminated glass slabs in pre- or post-breakage regime

Mechanical performance assessment of load-bearing elements under ordinary design actions is a priority when material degradation or even severe damage may be experienced. This is especially the case of unconventional constructional materials, or assembled systems like laminated glass (LG) solutions, in which multiple load-bearing components may suffer of different mechanical issues. This turns out in potential risk for customers, and in the need of specific design assumptions to address the problem. For load-bearing elements with direct interaction with final users (as pedestrian systems), a reciprocal modification of mechanical and dynamic parameters should be properly taken into account in the early design process. This paper explores the sensitivity of human-structure interaction effects and performance indicators for LG slab modules subjected to various walking paths of single pedestrians. An extended set of Finite Element (FE) nonlinear dynamic numerical simulations is presented, in which the variation of conventional performance indicators (such as acceleration peaks, deflection, vibration frequency) is investigated for two in-service LG modular units with intact (LGU) or partially fractured (LGF) glass layers. Walking scenarios are described with the support of consolidated deterministic method in use for pedestrian loads, in which walking features do not account for the features of structural background. The comparative analysis of numerical results with earlier experimental outcomes for two reference LG slab modules subjected to random walking conditions shows that time-domain nonlinear simulations with simplified pedestrian loads can efficiently capture some major effects of pedestrians, as well as the presence of damage in glass. At the same time, it is shown that fractured glass layers in LG pedestrian modules can still provide partial post-breakage contribution to the resisting section.