Empirical Peak Ground Velocity Attenuation Relations Based on Digital Broadband Records

Abstract eng:
Peak ground velocity (PGV) is very important to some large structures and buried pipelines. Previous PGV attenuation studies were mostly based on the integration to time histories of analog acceleration records which would inevitably introduce bias. In this paper we develop empirical attenuation relations for horizontal component of peak ground velocity in rock and soil sites. The strong motion datasets used to obtain the equations are recorded by Southern California Seismic Network. The digital broadband seismographes can directly record ground velocity and therefore can avoid errors of integration to acceleration records. 38 earthquakes with ML>5 from 1991 to 2001 were chosen, including 1992 Landers, 1994 Northridge, 1999 Hector earthquakes. The epicentral distance is up to 400 km. The site classification scheme was based on the averaged shear-wave velocity in upper 30m. 59 out of about 200 stations (i.e. approximate 30%) were grouped into rock site station. There are total 597 recordings for rock and 497 recordings for soil. An unweighted least-squares regression analysis was performed. Both horizontal components recorded at the same station were used independently. The PGV attenuation curves of rock and soil site have the similar trends for the same magnitude and distance, and the PGV on soil site is always larger. Comparisons with Huo(1989) was also illustrated, whose relations were calculated from velocity integrated from acceleration records. The curves have showed that for small earthquakes both equations have the same trends for rock, but he predict smaller values as the magnitude increases due to manually digitizing and low-cut filter. For soil, however, there are a significant discrepancy: his prediction decays more slowly than ours, leading to the prediction of our equation would be larger than Huo(1989) for the short distances, and smaller for the long distances, whether large magnitude or small.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket