Underground infrastructures and the importance of the people’s safety and structures robustness are relevant and contemporary issues in the civil engineering industry. The demand for this infrastructural typology has developed to such an extent that the need for a better and cost-beneficial knowledge of the risks is necessary. This thesis aims at bridging and developing further the knowledge on Fire safety engineering and structural engineering on the particular topic of spalling failure. Another objective of this thesis is the discussion over the possible replacement of the used procedures used to assess and guarantee tunnel safety. Both from a fire safety and structural engineering point of view, prescriptive measures and solutions are mostly proposed to accomplish a safe tunnel design. This causes the design to be non optimised and in some cases more costly than what is actually needed. From the fire safety side, the use of pre-given fire curves is put under discussion. Research has been conducted to asses which are the origins of the most widely used curves. Studies have also been performed to develop a practical analytical engineering method to analyse and estimate the consequences of a given fire scenario tailored to the specific tunnel under consideration. Subsequently the results of the analytical models have been compared with the results obtained with advanced Computational Fluid Dynamics tools. Finally, more complicated scenarios have been studied with the use of this software. On the other hand, from a structural engineering point of view, a new model able to describe the spalling mechanism has been proposed. The model predicts the spalling time for NSC elements and at the same time verifies which is the optimal thickness for the piece to spall. On top of that the possibilities for further use of the model in the description of the spalling mechanism for HSC and PPFRC elements have been investigated. Finally this two topics have been combined together and conclusion have been drawn.

To date, many fire accidents have occurred in tunnels all over the world. In the worst cases, there were catastrophic consequences with numerous human casualties and significant economic impacts. In the occurrence of such tunnel fires, besides the primary concern of saving human life, ensuring a certain level of structural integrity is also desired. Looking from the latter perspective, this thesis is focused on the fire-induced thermal exposure in concrete tunnels, aiming at coupling numerical simulations of idealized fire scenarios in computational fluid dynamics (CFD) software with the use of FEA for the analysis of structures, moving towards a performance-based approach in replacement of the currently adopted prescriptive approach. In this sense, fire imposes one of the most hazardous conditions for concrete structures, characterizing severe thermal degradation of the mechanical properties of the material. During fire exposure, steep thermal gradients are formed within the concrete structure. Those will always lead to the development of thermal stresses, the state of which being determined by several conditions (i.e. elastic modulus, thickness and supports). Besides the structural consequences of a fire, this thesis also deeply investigates the interface between a fire and the structure, which involves the interaction of three different modes of heat transfer, namely conduction, convection (related to the gas temperatures generated by a fire), and radiation (related to the radiation temperatures generated by a fire). While conduction within the solid phase regulates the amount of heat absorbed by the structure during a fire, radiation overcomes convection and becomes the primary mode of heat supply to the structure. The adiabatic surface temperature (AST) combines both the radiation and convection contributions to the total heat transfer to the structure into a single term, fully characterizing the fire phase in the structural analysis. Based on this concept, the differences between the two methodologies (i.e. prescriptive-based and performance-based) are exposed. From the prescriptive-based perspective, the development of the RWS curve and its indistinct use among protected and unprotected concrete tunnels is deeply discussed, together with the current fire resistance furnace tests. For the performance-based approach, the thermal exposure generated by two pool fire scenarios (200 MW and 30 MW) within the Piet Hein tunnel is addressed and the effects of longitudinal ventilation are highlighted.