地震动随机函数模型研究(Ⅱ)——参数统计与模型验证  

Research on random function model of strong ground motion(Ⅱ):parametric statistic and model certification

作　　者：安自辉[1] 李杰[1] 

机构地区：[1]同济大学建筑工程系,上海200092

出　　处：《地震工程与工程振动》2009年第6期40-47,共8页Earthquake Engineering and Engineering Dynamics

基　　金：国家自然科学基金委优秀创新研究群体科学基金项目(50321803)

摘　　要：依场地类别进行了强震记录分组,对模型参数的变化规律进行了统计分析。在模型随机参数向量满足独立性假设的前提下,得到了地震动随机函数模型的联合概率密度函数。引入数论选点方法对地震动随机函数模型的概率空间进行剖分,可以较少的样本点描述概率空间。以所选模型参数代表点代入地震动随机函数模型,即可以得到地震动时程样本集合。在集合层次上对比了模型预测地震动与真实记录的差异,两者在均值谱和标准差谱层次上均吻合较好,证实了模型预测结果的合理性。According to the site classification,the model parameters were divided into four classes,and the probability densities were estimated by the stochastic modeling method.Making the assumption about statistic independence,the joint probability density functions were obtained.In order to divide the probability-assigned space by a point set with cardinal number as small as possible,the number-theory method was introduced to determine the representative points of the model parameters.Substituting the represented points in the proposed random function model and performing the inverse Fourier transform yield the assembles of the stochastic accelerograms.The comparisons of the mean and standard deviation of the Fourier amplitude between the assembles of the predicted accelerograms and the counterparts of the recorded accelerograms were made,which showed that the prediction is reasonable.

关 键 词：地震动 随机函数 数论方法 概率空间 

分 类 号：P315.964[天文地球—地震学]