Abstract
A sequence of two strike-slip earthquakes occurred on April 14 and 16 (namely, the foreshock and the main shock), 2016 in Kumamoto Prefecture, Japan, and caused heavily damage to more than 7,000 wooden houses and residential buildings mainly in a central district of Mashiki town. The heavily damaged area of wooden houses was concentrated in a 0.3 km narrow belt that lies along the east-west direction (Fig. 1), where the belt is situated on a plain formed by several rivers and its geological setting could be a sedimentary deposit. This suggests that the concentration of building damage might be due to the local site effects in the area. There are two strong motion observation stations in Mashiki town. One is the KiK-net KMMH16, on the ground surface which is located north from the damage concentrated area. The other is the local government's intensity seismometer collaborated with JMA, called MTO in this paper, on the first floor of a building at Mashiki town office, which is located nearby the damage concentrated area and may be affected by soil-structure interaction (SSI) effects. In order to evaluate the local site effects on the seismic records, SSI effects on that should be clarified. Therefore, dynamic response analyses of the building at Mashiki town office are conducted for the earthquakes, considering SSI effects with nonlinearity based on geophysical and geotechnical surveys.
Mashiki town office is a three-story reinforced concrete building supported by pre-stressed concrete (PC) piles constructed in 1980 and has been retrofitted by out-flame in a longitudinal direction (Fig. 2). According to the visual inspections in post-earthquake damage evaluation, heavily damaged piles and some gaps caused between structure and soil were shown after the earthquakes (Photo1). The pseudo velocity response spectral value of acceleration records at the MTO is larger than that at the KMMH16 in a period range over 1 s during the main shock (Fig. 3, 4). In order to evaluate both effects of the local site and the SSI quantitatively, site investigations (i.e., borehole surveys, microtremor explorations and soil tests) were conducted (Fig. 5-9, Table1).
Then, the seismic responses of ground at the MTO are estimated by using the nonlinear seismic response analyses from the acceleration records on the ground surface at the KMMH16 (Fig. 10, 11). It is found from the analysis result that the acceleration response on the ground surface at the MTO is almost the same when compared to that at the KMMH16 (Fig. 12).
Dynamic response analyses are conducted by using nonlinear time history analyses to the SSI analysis model of the building (Fig. 13-15, Table1-2). A gap caused between pile and soil is modeled as slip behavior in a hysteresis rule of a soil spring at a shallower position. The foreshock and the main shock are continuously imposed on the analysis model. The SSI analyses have resulted in good agreement with the strong motion records observed at base foundation of the building during the earthquakes (Fig. 16). Furthermore, SSI effects include some nonlinearities of a pile-soil system at a shallower position, which is a gap caused between pile and soil and damage to pile head (Fig. 17).
The parametric studies indicate that the seismic response at the base foundation to the main shock which has a strong component in a period of 1 s is increased by a gap caused between pile and soil due to the foreshock or strength deterioration of pile head due to damage of PC pile, which contributes to a stiffness reduction of a pile-soil system and an increase in a base foundation response to an oscillation in the period of 1 s during the main shock (Table3, Fig. 18-20).
