A numerical study on seismic vulnerability of tunnel linings

The seismic vulnerability assessment of tunnels is generally based on empirical fragility curves (e.g. Hazus 1999; ALA 2001). They are derived from the statistical analysis of observed damages in past earthquakes. Some authors have calculated fragility curves based on the results of numerical analyses (e.g. Argyroudis and Pitilakis 2012), defining the damage index in terms of ratio between the flexural demand and capacity of the structural section of the tunnel lining. However, the tunnel lining is generally assumed as a continuous ring, while in the case of mechanized tunneling it is made of jointed segments. Recent studies modeling with FEA the non-linear dynamic interaction between the segmental lining and the soil (Fabozzi & Bilotta 2016) have shown the significant effect of the jointed pattern of a segmental lining. On one hand, it implies a lower structural demand than a continuous ring, due to its larger flexibility and compressibility in the transverse section; on the other hand, permanent rotations of the joints are often calculated at the end of the earthquake, that may represent a further structural fragility of a segmental lining, since they may produce dislocation of sealing gaskets among segments. Limited available post-earthquake reconnaissance (Dean et al., 2006) would seem to indicate that in some cases damage at the segments joints has been observed. In this study, a set of FEA was carried out using a strain hardening elastic-plastic model with small strain overlay (Benz et al., 2009). Several input signals were used and different ground conditions were modelled in order to define fragility curves for damage associated to joint permanent rotation in a segmental tunnel lining. The results are presented and discussed in the paper.