中文摘要：

地震过程中震动传播的局部场地效应通常用一维的场地反应分析方法来考虑,目前使用最多的是线性、等效线性化和非线性三类,其中国内外分别以LSSRLI-1和SHAKE2000为代表的等效线性化程序使用最为广泛。LSSRLI-1是我国地震安全评价程序,提出已有20多年,为我国的防震减灾事业做出了巨大贡献,但也在实践过程中表现出了一些缺点和不足,是否可以改进有待研究。本文以经典波动理论为基础编制了相应的土层反应分析线性计算程序,并与LSSRLI-1和SHAKE2000就硬和软两种场地上地表加速度反应谱和土体剪应变分布进行了对比分析,考虑了不同输入波以及不同震动强度对分析结果的影响,也讨论了土体剪应变计算偏差对地表反应谱的影响。结果表明：硬和软场地上SHAKE2000计算出地表加速度反应谱和土体剪应变分布均符合较好;硬和软场地上传递函数计算中LSSRLI-1、SHAKE2000和精确解三者一致;硬场地上LSSRLI-1计算所得土体剪应变分布同SHAKE2000和精确解相比偏差很小且对反应谱的影响可以忽略;某些情况下,软场地上LSSRLI-1计算出的土体剪应变分布同SHAKE2000和精确解相比明显偏大,该偏差会导致地表响应显著偏小,说明LSSRLI-1用于软场地地震反应分析时其剪应变求解方法有待改进。

英文摘要：

One-dimensional （ 1 D） site response analysis is commonly performed to account for local site effects on ground motion propagation during an earthquake. Available methods can be divided into linear, equivalent linear, and nonlinear, and in which the equivalent method represented by LSSRLI-1 and SHAKE2000 is the most widely used. It had been twenty more years since the safety evaluation program LSSRLI-1 was proposed, and it has made a great contribution to earthquake preparedness and disaster reduction for our country, and also some deficiencies were found in the practice. It need to be researched that if some improvements can be made for the evaluation program. On the basis of the classical wave motion theories, a linear program for seismic response analysis is cornpiled. When combined with the SHAKE2000 and LSSRLI-1 resuhs, comparisons for the ground surface acceleration spectra and the shear strain distribution are made. The influences of different incidence waves and the strength of shaking are completely considered. Analyses are made on the acceleration spectra differences induced by the shear strain variances. The result shows the ground response spectra and the shear strain distribution on hard and soft site, calculated by SHAKE2000, are both in good correspondence with the exact solution. The transform functions, calculated by LSSRLI-1, SHAKE2000 and exact solution, match well with each other. The shear strain differences on hard ground, when compared to SHAKE2000 and the exact solution, is really small, and the corre- sponding variances in the spectra acceleration can be ignored. In some cases, the shear strain on the soft site, calculated by LSSRLI-1, is obviously larger than the exact solution and SHAKE2000 results. The variances can bring extreme differences in the ground surface response. The method for computing the shear strain should be improved when the LSSRLI-1 is used for seismic response analysis on the soft sites.