Many buildings in Kobe City and its surrounding area suffered severe damage during the Hyogo-ken Nambu earthquake of 1995 in Japan. It is an essential point to clarify what characteristics the input motion had, and how the buildings responded, suffering damage or surviving the earthquake undamaged.
The objective of this paper, as the initial phase, is to grasp the global and quantitative seismic intensity to buildings in the heavily damaged area. First, we developed a relationship between calculated PGVs and damage ratio of residential houses in Chuo-ward, Kobe. Second, using the relationship, we evaluated the PGV distribution in the whole Hanshin area, and demonstrated that the evaluated values correspond to the peak velocity values of observed records in the evaluated area. Additionally, we demonstrated in other papers that the evaluated values were also corresponding to the overturning ratios of tombstones, and a two-dimensional PGV distribution in Nada-ward and Higashinada-ward evaluated by other researchers. Therefore, the evaluated values are considered to show one of the most credible seismic intensity distributions at present.
The evaluated PGVs in the most heavily damaged areas were estimated to be 110 cm/sec at least, which far exceeds the seismic force level supposed in the design stage. Nevertheless, few buildings designed by the current Building Code suffered severe damage. On the other hand, the peak velocity observed at the foundation of an office building near the heavily damaged area is about 85 cm/sec and is much less than PGVs in the heavily damaged area.
Then, it was supposed that effective input motion to buildings may be reduced by soil-structure interaction effects, like the filtering effects of foundation slabs. However, our PGV estimation at the site was nearly equal to the peak velocity of the record. Thus, we investigated the soil-structure interaction effects at the site. Response analyses of building models were carried out based on the soil profile, the foundation of the office building, and calculated ground motion at the site. Based on the results, we confirmed that soil-structure interaction effects were not so predominant and the PGV at the site should be much smaller than in the damage belt zone.