A Study of Selection and Scaling of Ground Motions for Structural Response History Analysis

• Main Authors: Tsai-Hsun Hsieh, 謝采洵
• Other Authors: 蔡克銓
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/72897955419709580124

碩士 === 國立臺灣大學 === 土木工程學研究所 === 100 === Nonlinear response history analysis (NLRHA) has become a frequently used procedure for the seismic demand evaluation for tall or irregular buildings. This study consists of two parts: ground motion scaling and ground motion selection. Bi-directional response history analyses (RHA) are adopted so that the torsional effects in the two-way asymmetric-plan buildings are included. This research investigates a multi-mode ground motion scaling (MMS) method in which the dynamic properties of the first few vibration modes are incorporated. The method considers the complete quadratic combination (CQC) rule in computing the scale factor and peak seismic demands in each of the two principal building axes. Multi-mode ground motion scaling method is divided into two approaches. One is the MMSV (base shear is chosen as the key design parameter) and the other is MMSD (roof displacement is chosen as the key design parameter). Using a 20-story two-way asymmetric-plan building, this study compares the effectiveness of the MMSV, MMSD with other common scaling procedures. It is illustrated that the MMS method is effective in reducing the scatter in the peak seismic demands computed from both the bi-directional RHA and response spectrum analysis (RSA). The MMSV method performs well in reducing the scatter of elastic shear force and inter-story drift demand estimates, while MMSD performs well in floor displacement and floor rotation demand estimates. For the inelastic building responses, MMSV provide rather good results in reducing the scatter of the peak seismic demands. For the NLRHA, it often requires the selection of input ground motions to represent the effects of the target response spectrum. In this study, the sum of squared errors (SSE) is used to determine the degree of the matching between the spectra. This study first applies multi-mode selection method, considers different weighting factors for different modes in calculating the SSE. It is demonstrated that MMSD (multi-mode selection method using roof displacement as the key design parameter) and common method perform well in spectrum matching and NLRHA for both 2-D and 3-D structural analyses. The second part of the research on ground motion selection is about the further modifications of the SSE. The SSEs are modified by weighting factors or normalization. It is found that the modified SSE using weighting factors or normalized shows little effect in both spectrum matching and reducing the scatter of the seismic response estimates.