Propagation of the earthquake motion towards the ground surface alters both the acceleration and frequency content of the motion. Acceleration–time record and Fourier amplitude spectrum of the motion reveal changes in the acceleration and frequency content. However, Fourier amplitude spectrum fails to give frequency–time variation. Wavelet transforms overcome this difficulty. In the current study, site response analysis of a liquefiable soil domain has been investigated employing wavelet transforms. Three earthquake motions with distinct predominant frequencies are considered. It is revealed that the moment soil undergoes initial liquefaction, it causes a spike in the acceleration–time history. From the analysis, frequency of the spikes is found to be greater than the predominant frequency of the acceleration time history recorded at the ground surface. Interestingly, the spikes belong to the sharp tips of the shear stress–shear strain curve. Immediately after the spike, acceleration deamplification is observed. Post-liquefaction deamplification (filtering) of the frequency components is also observed.

中文翻译：

---

采用连续小波变换的可液化土壤场地响应分析

地震运动向地表的传播改变了运动的加速度和频率内容。运动的加速度时间记录和傅里叶振幅谱揭示了加速度和频率内容的变化。然而，傅立叶幅度谱不能给出频率-时间变化。小波变换克服了这个困难。在目前的研究中，已经使用小波变换研究了可液化土壤域的场地响应分析。考虑了三种具有不同主要频率的地震运动。结果表明，在土壤经历初始液化的那一刻，它会导致加速时间历史的峰值。从分析来看，发现尖峰的频率大于在地面记录的加速度时间历史的主要频率。有趣的是，尖峰属于剪应力-剪应变曲线的尖端。在尖峰之后，立即观察到加速去放大。还观察到频率分量的液化后去放大（过滤）。