In the structural and earthquake engineering literature, probabilistic examination of all uncertainties and hazard levels considering the structural performance is suggested for design of existing buildings because of random nature of earthquakes. For different intensities of the earthquake, the fragility curves specify the achievement of the performance level in a probabilistic manner. For this purpose, some parameters such as the rotation and the axial deformation of the plastic hinge, the inter-story drift are usually selected as damage modes. Two factors affect the fragility curve: (1) damage from the probability function and (2) characteristics of the seismic scenarios such as Peak Ground Velocity (PGV), Peak Ground Velocity (PGV) and Peak Ground Acceleration (PGA). According to the derived probabilistic functions from intensity values for different limit states, Eq. (1) can be used as follows:
$$\:{F}_{i}\left(im\right)=p(D>{d}_{i}\left|IM=im\left)\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\right(1)\right.$$
where Fi(m) is the vulnerability probability and for earthquake intensity of IM = im, D is the damage level over a specific damage mode (di). PGA, PGV, or PGD can be selected as IM and different range of 0 to "n" can be defined for damage levels "i" as intact and damaged structures, respectively. In this paper, fragility curves were determined based on Shinozuka et al. [35], for Immediately Occupancy (IO), Life Safety (LS), and Collapse Prevention (CP) performance levels under critical single and successive shocks (foreshock, mainshock and aftershock). In these curves, the exceeding probability of the desired damage and PGA are introduced by vertical and horizontal axes, respectively. Proposed value of 0.005, 0.01, and 0.02 was selected for IO, LS, and CP performance levels, respectively [36–37]. Based on Figures (3) to (12), when the ratio of PGAm to PGAa ("m" and "a" present first shock and second shock) is greater than one, the exceeding probability of the desired damage increases. The comparison of median values shows that the PGA related to the given performance levels for consecutive shocks is less than singles. In other words, the exceeding probability of structural damage caused by successive shocks shocks is larger than singles, especially in short structures. By comparing the shear and flexural performance of the structures, the probability of failure in the LCF models with shear linked beams is lower than the structures with flexural linked beams. Also, as the number of stories increases, the possibility of failure has not necessarily increased or decreased.
6.1. Comparison of the probability of failure under seismic scenarios with and without seismic sequence
In this part, the probability of structural failure has been calculated for each model under single and consecutive records. As an example, for 5-story LCF frames with flexural linked beam (Fig. 6), the probability of failure for IO, LS and CP are 83%, 81% and 80% under single and 87%, 85% and 82% under successive records, respectively. It can be claimed that the probability of structural failure exposed to the seismic sequence phenomenon is higher than a single earthquakes. In other words, the proposed single design earthquake by the seismic design codes is not a suitable representative and the structures collapse earlier than the regulations assumptions. Because for all performance levels, the fragility curve for single shocks is lower than successive case. This means that the probability of exceeding the regulatory limits corresponding PGA for successive cases is lower than single case. Also, the probability of failure for 7-story LCF frames with flexural linked beams in IO, LS and CP are 78%, 75% and 73% under single shocks and 88%, 86% and 83% under successive shocks, respectively. Another example is for 11-story LCF frames with shear linked beam, which have the possibility of failure in IO, LS and CP are 76%, 73%, and 70% under single and 84%, 79%, and 73% under multiple records, respectively. The results indicate that the probability of failure in steel structures containing LCF exposed to consecutive accelerograms is higher than single shocks.
6.2. Comparison of the performance level under single and multiple accelerometers
In this section, in order to compare the performance level of the studied models under scenarios with and without seismic sequence for IO, LS and CP levels, the results of 5-story frame are shown in Fig. 13. In this figure, the vertical axis shows the obtained results from the performance level of the structure. According to this figure, in 5-story frame under successive records, the percentage of IO, LS and CP performance levels has increased by 4%, 4%, and 2% compared to the first shocks. The comparison of the performance level for 7 and 11 story frames is reported in Figs. 14 and 15, respectively. Based on the results of these figures, the percentage of IO, LS and CP performance levels has increased by 10%, 11%, and 8% compared to the single earthquakes. Meanwhile, the increase in results for the 11-story frame is equal to 12%, 9% and 17%, respectively.