Current seismic design provisions such as ASCE 7-10 provide criteria for selecting ground motions for conducting response history analysis. This study is the sequel of a companion paper (I – Ground Motion Selection) for assessment of the ASCE 7-10 criteria. To assess of the ASCE 7-10 criteria, nonlinear response history analyses of twelve single degree of freedom (SDF) systems and one multi-degree of freedom (MDF) system are conducted in this study. The results show that the target seismic demands for SDF can be predicted using the mean seismic demands over seven and ten ground motions selected according to the proposed method within an error of 30% and 20%, respectively

For estimating the seismic demand of buildings, most seismic design provisions permit conducting linear and nonlinear response history analysis. In order to obtain reliable results from response history analyses, a proper selection of input ground motions is required. In this study, an accurate algorithm for selecting and scaling ground motions is proposed, which satisfies the ASCE 7-10 criteria. In the proposed algorithm, a desired number of ground motions are sequentially scaled and selected from a ground motion library without iterations.

Currently, researches are being actively conducted in assessing seismic performance of nuclear facilities in USA and Europe. In particular, applying this technique of assessing seismic performance to design of isolation systems in nuclear power plants is being performed and then ASCE 4 Draft (2013) is being revised accordingly in the United States. In order to satisfy the probabilistic performance objectives described by seismic responses with certain confidence levels (ASCE 43, 2005), the probability distributions of these responses have to be defined. What is the minimum number of input ground-motions to obtain the probability distribution precise enough to represent the unknown actual distribution? Theoretical basis, for how to determine the minimum number of input ground-motions for given a logarithmic standard deviation to approximate the unknown actual median of the log-normal distribution within a range of error at a certain level of confidence, is introduced by Huang et al. (2008). However, the relationship between the level of confidence and the range of error is not stated in the previous study. In this paper, based on careful reviews on the previous work, the relationship between the level of confidence and the range of error is logically and explicitly stated. Furthermore, this relationship is also applied to derive the minimum number of input ground-motions in order to approximate the unknown actual logarithmic standard deviation. Several recommendations are made for determining the minimum number of input ground-motions in probabilistic assessment on seismic performance of facilities in nuclear power plants.

Structures in a nuclear power system are designed to be elastic even under an earthquake excitation. However a structural component such as an isolator shows inelastic behavior inherently. For the seismic assessment of nonlinear structures, response history analysis should be performed. In this study, the response of base isolation system was analyzed by response history analysis for the seismic performance assessment. Firstly, several seismic assessment criteria for a nuclear power plant structure were reviewed for the nonlinear response history analysis. Based on these criteria, the spectrum matched ground motion generation method modifying a seed earthquake ground motion time history was adjusted. Using these spectrum matched accelerograms, the distribution of displacement responses of the simplified base isolation system was evaluated. The resulting seismic responses excited by the modified ground motion time histories and the synthesized time history generated by stochastic approach were compared. And the response analysis of the base isolation system considering the different intensities in each orthogonal direction was performed.

비선형 구조계의 확률론적 지진해석 방법 중 대표적인 것은 지진 재해도 수준에 해당하는 입력지반운동 모델을 사용한 시간이력을 수행하여 그 응답의 확률분포를 예측하는 것이다. 이 연구에서는 널리 사용되고 있는 두 가지 입력지반운동 모델에 따른 구조계 응답의 분포특성 및 파손확률의 차이와 그 원인을 분석한다 입력지반운동 모델로는 실제 지진기록을 배율 조정하여 사용하는 배율조정 입력지반운동과 설계 응답스펙트럼에 상응하는 인공 지진기록을 사용하는 스펙트럼 맞춤 입력지반운동 두 가지를 고려한다. 동일한 지진재해도 수준을 고려한 해석결과 설계 응답스펙트럼에 상응하는 인공 지진기록을 사용한 입력지반운동 모델은 실제 지진기록을 배율 조정한 입력지반운동 모델보다 평균적으로 응답을 크게 평가하였고 이로 인해 파손확률 또한 더 큰 것으로 나타났다 이러한 경향은 연약한 지반에서 더욱 현저한 것으로 나타났다. 이러한 입력지반운동 모델에 따른 파손확률의 차이는 스펙트럼 맞춤 입력지반운동의 목표로 사용된 도로교 설계기준의 설계 응답스펙트럼이 실제 지진기록의 응답스펙트럼보다 장주기로 갈수록 응답을 크게 평가하도록 보수적으로 만들어졌기 때문인 것으로 나타났다.

The difference in phase and amplitude of ground motions recorded in different locations is generally known as spatial variation of seismic ground motions. This study focuses on the effects of spatial variation of earthquake ground motion on responses of adjacent buildings. The adjacent frame buildings are modeled considering soil-structure interaction (SSI) so that all the buildings can have interaction with each other under non-uniform ground motions. Based on Fast Fourier Transformation, spatially correlated non-uniform ground motions are generated compatible with known spectrum density function at different locations. Numerical analyses are carried out and the results are presented in terms of related parameters affecting the structural response using three different types of soil site classes. The results show that the effects of ground motion spatial variation are different for different site classes. The effect is more significant on rock site rather than clay site.