MDOF 해석을 통한 VS30 구간별 일반구조물의 PGA 증폭 모델 개발

Abstract

For estimating ground motion intensity measures on the surface from seismic sensors installed in structures, it is crucial to correct structural response effects embedded in the recorded signals. This study proposes a model for peak ground acceleration (PGA) amplification based on VS30, derived from multi-degree-of-freedom analysis. PGA amplification factor (AFPGA) is defined as the ratio of peak floor acceleration (PFA) of structures to PGA. The model includes three key input parameters: the natural period of the structures (Tn), the ratio of stories to the total number of stories in the structures, and the time-averaged shear wave velocity down to a depth of 30 meters. It is developed using 78 ground motion records from both domestic and international earthquakes. A LOESS smoothing technique is applied using 3 span values, with the optimal span of 0.1 is determined based on RMSE performance and an analysis of local trend characteristics in the dataset. The model is verified using empirical data from the CESMD global strong motion database, which includes classification by Tn into short, intermediate, and long periods. The results show that although the model tends to predict higher AFPGA values than those observed in real structures, it effectively reflects the overall amplification trends. This approach enables the pre-earthquake estimation of structural amplification, allowing for the use of seismic sensors installed in structures as a complementary monitoring network for seismic response.