PSI - Issue 78

Marco Civera et al. / Procedia Structural Integrity 78 (2026) 1783–1790

1787

4.2. Damage State 2 (DS2) In the second case, the (uncorrected) mean efficiency index stabilises around =0.38 after 1000 simulations (Figure 1.b). If one compares it to , , in this case, it shows that the RN efficiency has been reduced by circa 40% w.r.t. to its undamaged maximum , ; thus, less than what was obtained for DS1.

(a) (b) Fig. 1. Uncorrected road network (RN) efficiency for a scenario with return period =475 years and considering (a) DS1 and (b) DS2. As already discussed before, these results must be interpreted according to the intended use of this methodology. For the immediate post-event aftermath, it is plausible that any damage, even the smallest ones, can compromise the operability of the RN. Thus, the higher efficiency for DS1 is cautionary and should be used as proposed (but using Sc50) in (Miano et al., 2024). However, as discussed here, DS1 effects to assets (bridges and buildings) can be repaired or temporarily fixed in a matter of hours. Hence, the RN will be operative again sooner than after DS2 damages. If the corrective factor of 0.214, as proposed here, is applied to this last result, then , =0.081 becomes (sensibly) lower than the corresponding value as obtained for DS1 ( 0.281 ), highlighting that, in this instance, many more hours will be needed to recover and reopen the RN (here, it was assumed slightly more than 4.6 times, on average). 4.3. Damage State 3 (DS3) In this third case, the (uncorrected) mean efficiency index stabilises around =0.49 after the same amount of simulations (2000). This result is also consistent with the one reported in Section 4.2. of (Miano et al., 2024), shown here in Figure 2.a and 2.b, side-by-side for direct comparability. Please notice that, at this point, the probability of damage begins to become quite low. For the very first simulations, in no asset (bridge or building) the seismic demand exceeded the seismic capacity, thus causing to initially remain equal to , =0.64 . Nevertheless, as said, with enough simulations, the issue was solved. In uncorrected terms, once again, the efficiency of DS3 should be higher than when considering DS2, even if slightly so (a decrease of ~23% rather than ~30%). If corrected by multiplying with =0.049 , this is reduced to , =0.024 , following the decreasing trend seen before.