基于因子分析的地震动特征提取及潜在破坏势评估

基于多种地震动强度指标，采用因子分析提取地震动强度指标的特征因子，将提取的主因子进行线性组合，构造地震动多元强度主因子。以不同恢复力模型的单自由度以及多自由度体系为研究对象，基于2766条地震动记录，选取9个基本强度指标并采用因子分析提取其特征，同时对构造的地震动多元强度主因子与不同结构损伤指标进行相关性分析。研究结果表明:不同场地类别下各地震动强度指标所属的主因子趋于稳定；通过因子分析提取的两个主因子具备较好的可解释性；场地类别、强度折减系数和恢复力模型对构造的地震动多元强度主因子与结构损伤指标的相关性影响较小；构造的地震动多元强度主因子考虑了更多地震动信息，保持了其与不同结构损伤指标稳定且良好的相关性。

FACTOR ANALYSIS-BASED GROUND MOTION FEATURE EXTRACTION AND THE MEASUREMENT OF THE POTENTIAL STRUCTURAL DAMAGE

The characteristics of earthquake intensity measures (IMs) were preferentially extracted by factor analysis through considering multiple IMs, and the principal factor of multivariate IMs in terms of a linear combination of the extracted fundamental factors was then constructed. The nonlinear responses of single degree of freedom (SDOF) systems modelled with different hysteretic models and multiple degree of freedom (MDOF) systems subjected to 2766 ground motions records were used. Nine IMs were selected and their characteristics were extracted by factor analysis and the correlation between the principal factor of multivariate IMs and damage measures (DMs) was investigated. The results show that the IM components belong to a stable principal factor irrespective of site category and the two principal factors extracted by factor analysis have positive interpretability. The site category, hysteretic behavior and lateral strength ratio have little influence on the correlation coefficient between the principal factor of multivariate IMs and DMs. The principal factor of multivariate IMs shows significantly high-level correlation with DMs through considering multiple IMs.