中文摘要：

剖析了当今国际上频域等效线性化和时域非线性求解土层地震反应分析方法的典型代表SHAKE2000与DEEPSOIL，对目前其他学者FDM改进方式的可行性进行评判，采用新的直频法动剪模量阻尼比求解技术，提出新一代一维土层地震反应计算方法和程序SOILQUAKE，并用四种类别和巨厚场地实际记录检验了新程序。结果表明：SHAKE2000与DEEPSOIL在一类、二类场地上基本一致且结果可靠，三类和四类场地上两程序计算反应谱“矮粗胖”现象严重，与实际情况相差甚远；现有考虑频率相关性的FDM改进方式，其基本出发点存在定性错误；对硬场地和中硬场地，SOILQUAKE与SHAKE2000和DEEPSOIL结果大致相似，但在高频部分表现比SHAKE2000更加合理，与DEEPSOIL基本一致；对软场地和巨厚场地，S01LQUAKE克服了SHAKE2000和DEEPSOIL严重低估场地放大效应的缺点，模拟结果与实测记录较为接近，在软场地上优势显著，在巨厚软场地上优势十分显著。SOILQUAKE首先攻克现有软土层地震动计算结果“矮粗胖”和硬土中等效线性化方法缺少高频分量求解难点，具备了作为新一代一维土层地震反应计算方法的能力，将在http：／／www．soilquake．coin或http：／／www．soilquake．cn.上提供共享服务。

英文摘要：

The international typical seismic response evaluation techniques, SHAKE2000 and DEEPSOILwhich represent the equivalent linearization method in frequency domain and the nonlinear method in time domain, respectively, were analyzed. The feasibility of the existing improved frequency-dependent methods （FDM） were evaluated. Based on the direct frequency method for calculating the dynamic shear modulus and damping ratio, a new generation earthquake response analysis method and its code （SOILQUAKE） for one-dimensional soil layers were developed, and then they were examined based on the actual records of four types of sites and thick soft deposits. The research results indicate that for type I and type II sites, SHAKE2000 and DEEPSOIL may achieve basically similar and reliable results. But for type III and type IV sites, the response spectra of ground computed from the two programs obviously appear to be stocky, stub and plump, which may significantly deviate from the actual cases. The existing improved FDM techniques considering frequency dependence are based on the hypothesis with obvious qualitative error. For hard and moderate hard sites, SOILQUAKE, SHAKE2000 and DEEPSOIL may lead to similar results, but for high-frequency range, SOILQUAKE and DEEPSOIL may lead to similar results, which are more reasonable compared to those of SHAKE2000. For soft sites and thick soft deposits, SOILQUAKE may overcome the shortcoming associated with SHAKE2000 and DEEPSOIL which may severely underestimate the amplification effect of site, as its simulation results are closer to the measured records. The advantage of SOILQUAKE for soft sites is obvious, and for thick soft sites, its advantage is more remarkable. Using SOILQUAKE, the problem that the ground responses in soft sites obviously appear to be stocky, stub and plump and the problem that the equivalent linearization approaches lack the high frequency components in stiff sites can be obviated. Consequently, SOILQUAKE can be regarded as a new generation method for the calcu