Human stability during floods: Experimental tests on a physical model simulating human body

Urban floods are becoming more and more intense and frequent allover the world. Extreme events are the main triggering factors of such floods, and merit attention for what concerns the urban planning and emergency strategies. Numerical models aimed at investigating the optimal paths for evacuees escaping a flooded urban environment may be used by local authorities to properly understand how to improve people safety and mitigate the flood risk. Implementation of empirical laws in such models to describe the people stability in flooded areas is thus crucial to understand the behavior of evacuees and rescuers during emergency conditions. Laboratory experiments have been undertaken using a physical model representing a human body at quasi-natural scale, towed by an electrical engine in the water at rest. This represents a novel laboratory approach which exploits a non-inertial reference frame in motion with the model. The experimental results, obtained using different combinations of water depth and flow speed, have led to empirical laws which outline the stability conditions occurring when either the model front or the model back faces the flow, these respectively corresponding to Backward Toppling Instability (BTI) and Forward Toppling Instability (FTI). Such laws have been found through comparison with reference literature works, using various statistical methods. The FTI condition has been seen to largely improve the human stability compared to BTI, in contrast to the results of previous literature works, which stated an overall similarity between the results of the two toppling conditions. To better understand the role of the water flow during the different tests, hydraulic forces and moments have been measured. It has been seen that dynamic and static effects are comparable during high-speed conditions, especially due to a relevant fluid-model interaction and an increase of the water-surface level, while dynamic effects are negligible during low-speed conditions. The results of the present contribution can represent an important step forward for the numerical models applied to the framework of urban and emergency planning.